Mireya Diaz

Conducting indirect-treatment-comparison and network-meta-analysis studies: Report of the ISPOR task force on indirect treatment comparisons good research practices: Part 2

Evidence-based health care decision making requires comparison of all relevant competing interventions. In the absence of randomized controlled trials involving a direct comparison of all treatments of interest, indirect treatment comparisons and network meta-analysis provide useful evidence for judiciously selecting the best treatment(s). Mixed treatment comparisons, a special case of network meta-analysis, combine direct evidence and indirect evidence for particular pairwise comparisons, thereby synthesizing a greater share of the available evidence than traditional meta-analysis. This report from the International Society for Pharmacoeconomics and Outcomes Research Indirect Treatment Comparisons Good Research Practices Task Force provides guidance on technical aspects of conducting network meta-analyses (our use of this term includes most methods that involve meta-analysis in the context of a network of evidence). We start with a discussion of strategies for developing networks of evidence. Next we briefly review assumptions of network meta-analysis. Then we focus on the statistical analysis of the data: objectives, models (fixed-effects and random-effects), frequentist versus Bayesian approaches, and model validation. A checklist highlights key components of network meta-analysis, and substantial examples illustrate indirect treatment comparisons (both frequentist and Bayesian approaches) and network meta-analysis. A further section discusses eight key areas for future research.

Safety Profile of Robot-Assisted Radical Prostatectomy: A Standardized Report of Complications in 3317 Patients

BACKGROUND Previous studies attempting to assess complications after robot-assisted radical prostatectomy (RARP) are limited by their small numbers, short follow-up, or lack of risk factor analysis. OBJECTIVE To document complications after RARP by strict application of standardized reporting criteria. DESIGN, SETTING, AND PARTICIPANTS Between January 2005 and December 2009, 3317 consecutive patients underwent RARP at a tertiary referral center. Median follow-up was 24.2 mo (interquartile range: 12.4-36.9). INTERVENTION Transperitoneal RARP was performed by one of five surgeons-two experienced, three beginners. MEASUREMENTS Complications were captured by exhaustive review of multiple datasets, including our prospective prostate cancer database, claims data, and electronic medical and institutional morbidity and mortality records, and reported according to the Martin-Donat criteria. Complications were stratified by type (medical/surgical), Clavien classification, and timing of onset. Multivariable analysis of factors predictive of complications was performed. RESULTS AND LIMITATIONS The median hospitalization time was 1 d. There were 368 complications in 326 patients (9.8%), including a transfusion rate of 2.2%. We detected 79 medical complications in 78 patients (2.4%) and 289 surgical complications in 264 patients (8.0%). There were 242 minor (Clavien 1-2) and 126 major (Clavien 3-5) complications. Two hundred ninety-nine (81.3%) complications occurred within 30 d, 17 (4.6%) within 31-90 d, and 52 (14.1%) after 90 d from surgery. On multivariable analysis, preoperative prostate-specific antigen values and cardiac comorbidity were predictive for medical complications, whereas age, gastroesophageal reflux disease, and biopsy Gleason score were predictive of surgical complications. Limitations of this study include representing results from a single high-volume referral center and not including the learning curve of the two most experienced surgeons. CONCLUSIONS RARP is a safe operation, with an overall complication rate of 9.8%. Most complications occurred within 30 d of surgery.

Off-clamp Robot-assisted Partial Nephrectomy Preserves Renal Function: A Multi-institutional Propensity Score Analysis

BACKGROUND Ongoing efforts are focused on minimizing or eliminating renal ischemia during robot-assisted partial nephrectomy (RPN). Although various techniques allowing the elimination of renal hilar clamping have been described, large multi-institutional studies assessing perioperative and functional outcomes of this approach are lacking. OBJECTIVE To evaluate perioperative and functional outcomes of RPN without hilar clamping and to assess comparative effectiveness relative to clamped RPN. DESIGN, SETTING, AND PARTICIPANTS A multi-institutional data analysis of prospectively collected records of 886 RPNs performed by high-volume surgeons across five academic institutions between 2007 and 2011 was carried out. A total of 66 patients who underwent RPN without hilar clamping were identified. After the exclusion of 17 patients, perioperative results of 49 patients were compared against propensity score matched clamped controls. INTERVENTION RPN without hilar clamping. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS Descriptive statistics and propensity score matching. RESULTS AND LIMITATIONS Patients undergoing off-clamp RPN had a mean tumor size of 2.5 cm (standard deviation [SD]: ± 2.1) and a mean RENAL nephrometry score of 5.3 (SD: ± 1.5). The mean preoperative estimated glomerular filtration rate (eGFR) was 81 (SD: ± 29). The mean estimated blood loss (EBL) was 210 ml (SD: ± 212), and the mean operative time was 155 min (SD: ± 46). No Clavien 3-5 complications were recorded. The mean postoperative change in eGFR was 3% at first follow-up (1-3 mo), and no patient required postoperative dialysis. The positive surgical margin rate was 3% (n=2), with no disease recurrence reported at a mean follow-up of 21 mo. In propensity score matched analyses, the off-clamp RPN patients had a significantly shorter mean operative time (156 min compared with 185 min, p<0.001), a higher EBL (228 ml compared with 157 ml, p=0.009), and a smaller decrease in eGFR (2% compared with -6%, p=0.008). The retrospective analysis was the main limitation of this study. CONCLUSIONS With appropriately selected patients and adequate surgeon experience, off-clamp RPN is safe and feasible. Off-clamp RPN was associated with higher EBL, shorter operative times, and smaller decrease in renal function.

Effect of metabolic syndrome on pathologic features of prostate cancer

OBJECTIVE The prevalence of metabolic syndrome has been increasing worldwide, however its association with prostate cancer (CaP) is unclear. We reviewed patients undergoing robot assisted radical prostatectomy (RARP) to evaluate if those with metabolic syndrome had more aggressive disease. MATERIALS AND METHODS A prospective database of patients undergoing RARP between January 2005 and December 2008 (n = 2756) was queried for components of metabolic syndrome (BMI ≥ 30 and ≥ 2 of the following: hypertension, diabetes or elevated blood glucose, and dyslipidemia; n = 357). Patients with no components of metabolic syndrome were used as controls (n = 694). Biopsy and final pathology were compared between the 2 groups using all controls, and using best-matched controls (n = 357) based on greedy matching by propensity score. RESULTS Compared with unmatched controls, metabolic syndrome patients had higher pathology Gleason grade (≥ 7: 78% vs. 64%, P < 0.001) and higher pathologic stage (≥ T3 disease: 43% vs. 31%, P < 0.001). After controlling for confounders, those with metabolic syndrome when compared with best-matched controls had maintained the greater pathology Gleason grade (≥ 7: 78% vs. 64%, P < 0.001) and pathologic stage (≥ T3 disease: 43% vs. 32%, P < 0.001). They also had significantly greater pathologic upgrading of Gleason grade 6 adenocarcinoma found on biopsy compared with best-matched controls (63% vs. 45%, P < 0.001). On pathology, a 2-fold increase in Gleason 8 and greater was noted between patients with metabolic syndrome and best-matched controls (15% vs. 8%). CONCLUSIONS After controlling for confounders, patients with metabolic syndrome were found to have higher Gleason grade and tumor stage on final pathology and were more likely to have upgrading.

Enuretic children with obstructive sleep apnea syndrome: Should they see otolaryngology first?

OBJECTIVES To study: (1) the prevalence of diurnal urinary incontinence (DI) and nocturnal enuresis (NE) in children with obstructive sleep apnea syndrome (OSAS) who underwent surgery for their upper airway symptoms, (2) the postoperative rate of enuresis resolution, and (3) factors that may predict lack of improvement post surgery. PATIENTS AND METHODS An observational, pilot study of children 5-18 years of age with OSAS and NE who underwent tonsillectomy and/or adenoidectomy (T&A) between 2008 and 2010 was performed. Study consisted of a phone interview and chart review. Severity of NE and DI, frequency, arousal and sleeping disturbances were assessed pre and post T&A. Factors associated with failure to respond were analyzed using a logistic regression model. RESULTS Among the 417 children who underwent T&A, 101 (24%) had NE (61 males, mean age 7.8 ± 2.5 years), and of these 24 had associated DI (6%). Mean postoperative follow-up was 11.7 months. Of the 49 whose NE responded to T&A (49%), 30 resolved within 1 month postoperatively. DI resolved in 4 children (17%). There was a statistically significant difference between responders and non-responders regarding the presence of prematurity, obesity, family history of NE, type of enuresis, enuresis severity, and ability to be easily aroused. CONCLUSION NE was present in about one fourth of children with OSAS undergoing surgery, and resolved in about half. Lower response rate was associated with prematurity, obesity, family history of NE, presence of non-monosymptomatic NE, severe NE preoperatively, and arousal difficulties.

Long-term follow-up of patients undergoing percutaneous suprapubic tube drainage after robot-assisted radical prostatectomy (RARP)

Study Type – Therapy (case series)

Application of the Statistical Process Control Method for Prospective Patient Safety Monitoring During the Learning Phase: Robotic Kidney Transplantation with Regional Hypothermia (IDEAL Phase 2a–b)

BACKGROUND Traditional evaluation of the learning curve (LC) of an operation has been retrospective. Furthermore, LC analysis does not permit patient safety monitoring. OBJECTIVES To prospectively monitor patient safety during the learning phase of robotic kidney transplantation (RKT) and determine when it could be considered learned using the techniques of statistical process control (SPC). DESIGN, SETTING AND PARTICIPANTS From January through May 2013, 41 patients with end-stage renal disease underwent RKT with regional hypothermia at one of two tertiary referral centers adopting RKT. Transplant recipients were classified into three groups based on the robotic training and kidney transplant experience of the surgeons: group 1, robot trained with limited kidney transplant experience (n=7); group 2, robot trained and kidney transplant experienced (n=20); and group 3, kidney transplant experienced with limited robot training (n=14). INTERVENTION We employed prospective monitoring using SPC techniques, including cumulative summation (CUSUM) and Shewhart control charts, to perform LC analysis and patient safety monitoring, respectively. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS Outcomes assessed included post-transplant graft function and measures of surgical process (anastomotic and ischemic times). CUSUM and Shewhart control charts are time trend analytic techniques that allow comparative assessment of outcomes following a new intervention (RKT) relative to those achieved with established techniques (open kidney transplant; target value) in a prospective fashion. RESULTS AND LIMITATIONS CUSUM analysis revealed an initial learning phase for group 3, whereas groups 1 and 2 had no to minimal learning time. The learning phase for group 3 varied depending on the parameter assessed. Shewhart control charts demonstrated no compromise in functional outcomes for groups 1 and 2. Graft function was compromised in one patient in group 3 (p<0.05) secondary to reasons unrelated to RKT. In multivariable analysis, robot training was significantly associated with improved task-completion times (p<0.01). Graft function was not adversely affected by either the lack of robotic training (p=0.22) or kidney transplant experience (p=0.72). CONCLUSIONS The LC and patient safety of a new surgical technique can be assessed prospectively using CUSUM and Shewhart control chart analytic techniques. These methods allow determination of the duration of mentorship and identification of adverse events in a timely manner. A new operation can be considered learned when outcomes achieved with the new intervention are at par with outcomes following established techniques. PATIENT SUMMARY Statistical process control techniques allowed for robust, objective, and prospective monitoring of robotic kidney transplantation and can similarly be applied to other new interventions during the introduction and adoption phase.

Robotic partial nephrectomy in the setting of prior abdominal surgery

Study Type – Therapy (case series)

Contemporary Nationwide Patterns of Self-reported Prostate-Specific Antigen Screening

Funding/Support: Dr Smith is supported by National Institutes of Health/ National Heart, Lung, and Blood Institute training grant T32HL076139. Dr Weiss is supported by National Institutes of Health/National Heart, Lung, and Blood Institute grant K23HL118139 and a grant from the Parker B. Francis Fellowship Program. Dr Wunderink is supported in part by Centers for Disease Control and Prevention grant 1U18IP000490.

Findings in 12-core transrectal ultrasound-guided prostate needle biopsy that predict more advanced cancer at prostatectomy: analysis of 388 biopsy-prostatectomy pairs.

We analyzed 5 features on 12-core transrectal ultrasound-guided prostate needle biopsy (TRUS) to predict the extent of cancer at radical prostatectomy (RP). In 388 TRUS-RP pairs, number of positive cores (NPC), percentage of each core involved (%PC), perineural invasion (PNI), Gleason score (GS), distribution of positive cores (DPC), and preoperative prostate-specific antigen (PSA) were correlated with extraprostatic extension (EPE), seminal vesicle invasion (SVI), positive surgical margin (R1), positive lymph nodes (N1), and tumor volume. All features predicted EPE and SVI. NPC, GS, %PC, and PNI strongly predicted R1 status. RP tumor volume was directly proportional to the NPC and %PC. PSA alone and with selected biopsy findings correlated with tumor volume, stage, SVI, and N1 (P < .0001). Contiguous DPC was a significant risk for EPE and SVI (P < .0001) compared with isolated positive cores. Findings at 12-core TRUS along with preoperative PSA reliably predict advanced local disease and have practical value as guides to effective planning for surgical resections.