Benign hysterectomy performed by gynecologic oncologists: Is selection bias altering our ability to measure surgical quality?

OBJECTIVE: To compare the characteristics of women undergoing hysterectomy for benign disease with either a benign gynecologist or a gynecologic oncologist and to assess for differences in complication rates with and without risk adjustment. METHODS: Patients undergoing benign hysterectomy recorded in the National Surgical Quality Improvement Program (NSQIP) targeted hysterectomy file in 2015 were identified. The primary outcome was any postoperative complication. Stratified analysis was performed by route of surgery. Bivariable tests and modified Poisson regression were used to adjust for confounding by procedure type and patient characteristics. RESULTS: We identified 17,639 patients who underwent hysterectomy for benign pathology, primary surgeon was a benign gynecologist (82%) or gynecologic oncologist (18%). Patients who underwent surgery with gynecologic oncologists were older (51yo v 46yo), had a higher mean BMI (32 v 30), and a higher prevalence of prior abdominal surgery (29% v 25%, p < 0.001), diabetes (10.6% v 7.0%), hypertension (34% v 25%) and higher ASA and Charlson comorbidity scores (p < 0.001, for all). For laparoscopy, surgery with a gynecologic oncologist was associated with a decreased risk of complication (RR 0.80, 95% CI 0.66-0.98). For laparotomy, surgery with a gynecologic oncologist was associated with an increased risk of complication (RR 1.18 95% CI 1.01-1.38), however, this was no longer the case with risk adjustment (aRR 0.90, 95% CI 0.76-1.07). CONCLUSIONS: Patients operated on by gynecologic oncologists have a higher prevalence of risk factors for complication compared to those operated on by benign gynecologists even with a benign indication for surgery. Quality measurement should account for this selection bias.

Referral Bias in Primary Total Knee Arthroplasty: Retrospective Analysis of 22,614 Surgeries in a Tertiary Referral Center.

BACKGROUND: Patients who travel a significant distance to obtain surgical treatment typically experience better outcomes. This is called the referral bias and can limit the generalizability of studies performed at large tertiary care centers. We explored the influence of referral bias by comparing the clinical characteristics and outcomes of total knee arthroplasty (TKA) at a large tertiary care hospital in the United States. METHODS: The study cohort included 22,614 primary TKA procedures performed between 1985 and 2010. Patients were stratified into 5 groups using home address zip codes and according to travel distance from the hospital. Clinical characteristics and the risk of TKA complications and surgical outcomes (instability, surgical-site infections, and thrombovascular complications within the first year, reoperations, revisions, and mortality) were compared across the 5 groups. RESULTS: Compared with local patients, patients who traveled from other parts of the United States were significantly younger (mean age 67.8 vs 68.5 years; P < .05), were more likely to be male (47% vs 38%, P < .001), had lower body mass index (mean 30.4 vs 31.8 kg/m2; P < .001), were more likely to have inflammatory arthritis or neoplasms as surgical indications (P < .05), and were more likely to have a history of prior surgeries on the same knee (20% vs 14%; P < .001). Referral patients also had significantly higher American Society of Anesthesiologists scores and longer operative times (mean 173 vs 156 minutes P < .001). Despite these differences, the risk of instability, surgical-site infections, thrombovascular complications, reoperations, and revision surgeries were similar across the 5 groups. CONCLUSION: Although referral patients differ from local patients, the groups seem to experience largely similar complication and revision rates after TKA.

Balancing Contamination and Referral Bias in a Randomized Clinical Trial: An Application of Pseudo-Cluster Randomization.

In randomized trials of provider-focused clinical interventions, treatment allocation often cannot be blinded to participants, study staff, or providers. The choice of unit of randomization (patient, provider, or clinic) entails tradeoffs in cost, power, and bias. Provider- or clinic-level randomization can minimize contamination, but it incurs the equally problematic potential for referral bias; that is, because arm assignment of future participants generally cannot be concealed, differences between arms may arise in the types of patients enrolled. Pseudo-cluster randomization is a novel study design that balances these competing validity threats. Providers are randomly assigned to an imbalanced proportion of intervention-arm participants (e.g., 80% or 20%). Providers can be masked to the imbalance, avoiding referral bias. Contamination is reduced because only a minority of control-arm participants are treated by majority-intervention providers. Pseudo-cluster randomization was implemented in a randomized trial of a decision support intervention to manage depression among patients receiving human immunodeficiency virus care in the southern United States in 2010-2014. The design appears successful in avoiding referral bias (participants were comparable between arms on important characteristics) and contamination (key depression treatment indicators were comparable between usual care participants managed by majority-intervention and majority-usual care providers and were markedly different compared with intervention participants).

Is there referral bias in outcomes of septal myectomy for hypertrophic cardiomyopathy?

PURPOSE: To determine the potential impact of referral bias on short- and long-term outcomes following septal myectomy for hypertrophic cardiomyopathy. METHODS: We reviewed 2303 adult patients who underwent transaortic septal myectomy for obstructive hypertrophic cardiomyopathy from January 1993 to April 2016. Patients were divided into 3 groups according to their permanent address: local (state) residents (n = 324), regional (surrounding 5 states) patients (n = 515), and national (outside 5 states) patients (n = 1464). RESULTS: Patient groups were similar for age, sex, preoperative New York Heart Association class, and left ventricular ejection fraction. Local patients had increased prevalence of diabetes mellitus (13%, 11%, 8%; P = .006), coronary artery disease (25%, 21%, 19%; P = .031), severe chronic lung disease (2.3%, 1.9%, 0.4%; P < .001), and atrial fibrillation (24%, 18%, 19%; P = .045) when compared with regional and national patients. Echocardiographic features did not differ between the 3 groups, including prevalence of moderate or greater mitral regurgitation (59%, 61%, 56%; P = .161). Local and regional patients were more likely to undergo concomitant procedures than national patients (P < .001). Mitral valve surgery was performed in 9.6% of the patients, more commonly in local and regional patients (12%, 12%, 8%; P = .018). There were 11 operative deaths (0.5%), and early mortality was similar among the groups. Geographic origin did not impact overall late survival. CONCLUSIONS: Compared with distant referrals, local patients who undergo septal myectomy at our institution have more comorbid conditions, and require more concomitant surgical procedures. Despite these differences, referral patterns did not impact early or late outcomes following transaortic septal myectomy.

Impact of referral bias on prognostic studies outcomes: insights from a population-based cohort study on infective endocarditis.

PURPOSE: Prognostic studies derived from samples of patients managed in tertiary hospitals are subject to referral bias. We aimed to characterize this bias using the example of infective endocarditis. METHODS: We analyzed data from a French population-based cohort, which included 497 patients with infective endocarditis. Patients were admitted directly to a tertiary hospital (Group T), admitted to a non-tertiary hospital and referred to a tertiary hospital (Group NTT) or not (Group NT). We compared patients' characteristics, survival rates and prognostic factors between groups. RESULTS: Compared with Group T (n = 291), NTT patients (n = 144) were more often males (81.3% vs. 72.5%; P = .046), injection drug users (9.7% vs. 4.5%; P = .033), and had more frequent surgical indications (78.5% vs. 64.3%; P = .003). Compared with Group NT (n = 62), NTT patients were more often males (81.3% vs. 67.7%; P = .034) and had surgical indications more often (78.5% vs. 19.4%; P < .001). One-year survival was higher in NTT + T patients than in NT patients (73.0% vs. 56.1%; P = .01). Prognostic factors and hazard ratios estimates varied across groups. CONCLUSIONS: When derived from samples mixing patients admitted directly and those referred to tertiary hospitals, validity of characteristics description, survival estimates, and hazard ratios is threatened by referral bias.

Data on prognostic factors associated with 3-month and 1-year mortality from infective endocarditis.

This article describes supplementary tables and figures associated with the research paper entitled "Impact of referral bias on prognostic studies outcomes: insights from a population-based cohort study on infective endocarditis". The aforementioned paper is a secondary analysis of data from the EI 2008 cohort on infective endocarditis and aimed at characterising referral bias. A total of 497 patients diagnosed with definite infective endocarditis between January 1st and December 31st 2008 were included in EI 2008. Data were collected from hospital medical records by trained clinical research assistants. Patients were divided into three groups: admitted to a tertiary hospital (group T), admitted to a non-tertiary hospital and referred secondarily to a tertiary hospital (group NTT) or admitted to a non-tertiary hospital and not referred (group NT). The pooled (NTT+T) group mimicked studies recruiting patients in tertiary hospitals only. Two different starting points were considered for follow up: date of first hospital admission and date of first admission to a tertiary hospital if any (hereinafter referred to as "referral time"). Referral bias is a type of selection bias which can occur due to recruitment of patients in tertiary hospitals only (excluding those who are admitted to non-tertiary hospitals and not referred to tertiary hospitals). This bias may impact the description of patients' characteristics, survival estimates as well as prognostic factors identification. The six tables presented in this paper illustrate how patients' selection (population-based sample [pooled (NT+NTT+T) group] versus recruitment in tertiary hospitals only [pooled (NTT+T) group]) might impact Hazards Ratios values for prognostic factors. Crude and adjusted Cox regression analyses were first performed to identify prognostic factors associated with 3-month and 1-year mortality in the whole sample using inclusion as the starting point. Analyses were then performed in the pooled (NTT+T) group first using inclusion as the starting point and finally using referral time as the starting point. Figures 1 to 3 illustrate how HR increase with time for covariates that were considered as time-varying covariates (covariate*time interaction).

Age, Sex, and Gene Expression Score identifies a symptomatic, nondiabetic male patient as being at high risk of obstructive coronary artery disease.

In October 2015, a 74-year-old Caucasian male patient (past medical history of hyperlipidemia, paroxysmal atrial fibrillation, hypertension, and hypothyroidism) presented to the cardiologist for follow-up outpatient evaluation of exertional chest pain. The patient had recently been seen at the Emergency Department for the same complaint. At that time, the patient's cardiac markers, EKG, and pharmacological nuclear stress testing were all reported as normal. At presentation to the cardiologist, the patient's physical examination findings were unremarkable. Over the course of the following year, repeat electrocardiograms and myocardial perfusion imaging studies demonstrated no evidence of ischemia. Despite the persistence of symptoms, the patient was reluctant to undergo invasive testing. The cardiologist ordered a simple blood test: the Age, Sex, and Gene Expression Score, which provides the current likelihood of obstructive coronary artery disease in nondiabetic patients. Based on the high Age, Sex, and Gene Expression Score result, the patient underwent invasive coronary angiography and a 98% stenotic lesion in the proximal left anterior descending artery was discovered. A drug-eluting coronary stent was placed and resulted in the complete resolution of the patient's symptoms.

Criteria, prevalence, and phenotypes of polycystic ovary syndrome.

Polycystic ovary syndrome (PCOS) is a highly prevalent disorder effecting reproductive-aged women worldwide. This article addresses the evolution of the criteria used to diagnosis PCOS; reviews recent advances in the phenotypic approach, specifically in the context of the extended Rotterdam criteria; discusses limitations of the current criteria used to diagnosis, particularly when studying adolescents and women in the peri- and postmenopause; and describes significant strides made in understanding the epidemiology of PCOS. This review recognizes that although there is a high prevalence of PCOS, there is increased variability when using Rotterdam 2003 criteria, owing to limitations in population sampling and approaches used to define PCOS phenotypes. Last, we discuss the distribution of PCOS phenotypes, their morbidity, and the role that referral bias plays in the epidemiology of this syndrome.

Phenotypes and body mass in women with polycystic ovary syndrome identified in referral versus unselected populations: systematic review and meta-analysis.

OBJECTIVE: To compare the prevalence of polycystic ovary syndrome (PCOS) phenotypes and obesity among patients detected in referral versus unselected populations. DESIGN: Systematic review and meta-analysis. SETTING: Not applicable. PATIENT(S): Thirteen thousand seven hundred ninety-six reproductive-age patients with PCOS, as defined by the extended Rotterdam 2003 criteria. INTERVENTION(S): Review of PUBMED, EMBASE, and Cochrane Library, 2003-2016. Only observational studies were included. Data were extracted using a web-based, piloted form and combined for meta-analysis. MAIN OUTCOME MEASURE(S): PCOS phenotypes were classified as follows: phenotype A, clinical and/or biochemical hyperandrogenism (HA) + oligo-/anovulation (OA) + polycystic ovarian morphology (PCOM); phenotype B, HA+OA; phenotype C, HA+PCOM; and phenotype D, OA+PCOM. RESULT(S): Forty-one eligible studies, reporting on 43 populations, were identified. Pooled estimates of detected PCOS phenotype prevalence were consequently documented in referral versus unselected populations, as [1] phenotype A, 50% (95% confidence interval [CI], 46%-54%) versus 19% (95% CI, 13%-27%); [2] phenotype B, 13% (95% CI, 11%-17%) versus 25% (95% CI, 15%-37%); [3] phenotype C, 14% (95% CI, 12%-16%) versus 34% (95% CI, 25-46%); and [4] phenotype D, 17% (95% CI, 13%-22%) versus 19% (95% CI, 14%-25%). Differences between referral and unselected populations were statistically significant for phenotypes A, B, and C. Referral PCOS subjects had a greater mean body mass index (BMI) than local controls, a difference that was not apparent in unselected PCOS. CONCLUSION(S): The prevalence of more complete phenotypes in PCOS and mean BMI were higher in subjects identified in referral versus unselected populations, suggesting the presence of significant referral bias.

Epidemiology of infective endocarditis in a large Belgian non-referral hospital.

OBJECTIVES: Guidelines for diagnosis of infective endocarditis are largely based upon epidemiological studies in referral hospitals. Referral bias, however, might impair the validity of guidelines in non-referral hospitals. Recent studies in non-referral care centres on infective endocarditis are sparse. We conducted a retrospective epidemiological study on infective endocarditis in a large non-referral hospital in a Belgian city (Kortrijk). METHODS: The medical record system was searched for all cases tagged with a putative diagnosis of infective endocarditis in the period 2003-2010. The cases that fulfilled the modified Duke criteria for probable or definite infective endocarditis were included. RESULTS: Compared to referral centres, an older population with infective endocarditis, and fewer predisposing cardiac factors and catheter-related infective endocarditis is seen in our population. Our patients have fewer prosthetic valve endocarditis as well as fewer staphylococcal endocarditis. Our patients undergo less surgery, although mortality rate seems to be highly comparable with referral centres, with nosocomial infective endocarditis as an independent predictor of mortality. CONCLUSION: The present study suggests that characteristics of infective endocarditis as well as associative factors might differ among non-referral hospitals and referral hospitals.

The conundrum of primary snoring in children: what are we missing in regards to cognitive and behavioural morbidity?

Sleep disordered breathing (SDB) is common in children and describes a continuum of nocturnal respiratory disturbance from primary snoring (PS) to obstructive sleep apnoea (OSA). Historically, PS has been considered benign, however there is growing evidence that children with PS exhibit cognitive and behavioural deficits equivalent to children with OSA. There are two popular mechanistic theories linking SDB with daytime morbidity: hypoxic insult to the developing brain; and sleep disruption due to repeated arousals. These theories apply well to OSA, but children with PS experience neither hypoxia nor increased arousals when compared to non snoring controls. So what are we missing? This review summarises the literature examining daytime morbidity in children with PS and discusses the current debates surrounding this relationship. Specifically, questions exist as to the sensitivity of our standard assessment techniques to measure subtle hypoxia and arousal. There is also a suggestion that the association between PS and daytime morbidity may not be mediated by nocturnal respiratory disturbance at all, but by a number of other comorbid, but perhaps unrelated factors. As approximately 70% of children with SDB are diagnosed with PS, but are rarely treated, a paradigm shift in the investigation of PS may be required.