Outlier analysis for accelerating clinical discovery: An augmented intelligence framework and a systematic review.

Clinical discoveries largely depend on dedicated clinicians and scientists to identify and pursue unique and unusual clinical encounters with patients and communicate these through case reports and case series. This process has remained essentially unchanged throughout the history of modern medicine. However, these traditional methods are inefficient, especially considering the modern-day availability of health-related data and the sophistication of computer processing. Outlier analysis has been used in various fields to uncover unique observations, including fraud detection in finance and quality control in manufacturing. We propose that clinical discovery can be formulated as an outlier problem within an augmented intelligence framework to be implemented on any health-related data. Such an augmented intelligence approach would accelerate the identification and pursuit of clinical discoveries, advancing our medical knowledge and uncovering new therapies and management approaches. We define clinical discoveries as contextual outliers measured through an information-based approach and with a novelty-based root cause. Our augmented intelligence framework has five steps: define a patient population with a desired clinical outcome, build a predictive model, identify outliers through appropriate measures, investigate outliers through domain content experts, and generate scientific hypotheses. Recognizing that the field of obstetrics can particularly benefit from this approach, as it is traditionally neglected in commercial research, we conducted a systematic review to explore how outlier analysis is implemented in obstetric research. We identified two obstetrics-related studies that assessed outliers at an aggregate level for purposes outside of clinical discovery. Our findings indicate that using outlier analysis in clinical research in obstetrics and clinical research, in general, requires further development.

Augmented Intelligence for Clinical Discovery in Hypertensive Disorders of Pregnancy Using Outlier Analysis.

Objectives Clinical discoveries are heralded by observing unique and unusual clinical cases. The effort of identifying such cases rests on the shoulders of busy clinicians. We assess the feasibility and applicability of an augmented intelligence framework to accelerate the rate of clinical discovery in preeclampsia and hypertensive disorders of pregnancy-an area that has seen little change in its clinical management. Methods We conducted a retrospective exploratory outlier analysis of participants enrolled in the folic acid clinical trial (FACT, N=2,301) and the Ottawa and Kingston birth cohort (OaK, N=8,085). We applied two outlier analysis methods: extreme misclassification contextual outlier and isolation forest point outlier. The extreme misclassification contextual outlier is based on a random forest predictive model for the outcome of preeclampsia in FACT and hypertensive disorder of pregnancy in OaK. We defined outliers in the extreme misclassification approach as mislabelled observations with a confidence level of more than 90%. Within the isolation forest approach, we defined outliers as observations with an average path length z score less or equal to -3, or more or equal to 3. Content experts reviewed the identified outliers and determined if they represented a potential novelty that could conceivably lead to a clinical discovery. Results In the FACT study, we identified 19 outliers using the isolation forest algorithm and 13 outliers using the random forest extreme misclassification approach. We determined that three (15.8%) and 10 (76.9%) were potential novelties, respectively. Out of 8,085 participants in the OaK study, we identified 172 outliers using the isolation forest algorithm and 98 outliers using the random forest extreme misclassification approach; four (2.3%) and 32 (32.7%), respectively, were potential novelties. Overall, the outlier analysis part of the augmented intelligence framework identified a total of 302 outliers. These were subsequently reviewed by content experts, representing the human part of the augmented intelligence framework. The clinical review determined that 49 of the 302 outliers represented potential novelties.  Conclusions Augmented intelligence using extreme misclassification outlier analysis is a feasible and applicable approach for accelerating the rate of clinical discoveries. The use of an extreme misclassification contextual outlier analysis approach has resulted in a higher proportion of potential novelties than using the more traditional point outlier isolation forest approach. This finding was consistent in both the clinical trial and real-world cohort study data. Using augmented intelligence through outlier analysis has the potential to speed up the process of identifying potential clinical discoveries. This approach can be replicated across clinical disciplines and could exist within electronic medical records systems to automatically identify outliers within clinical notes to clinical experts.

Do Case Reports and Case Series Generate Clinical Discoveries About Preeclampsia? A Systematic Review.

Background: Preeclampsia is a leading cause of maternal and perinatal mortality and morbidity. The management of preeclampsia has not changed much in more than two decades, and its aetiology is still not fully understood. Case reports and case series have traditionally been used to communicate new knowledge about existing conditions. Whether this is true for preeclampsia is not known. Objective: To determine whether recent case reports or case series have generated new knowledge and clinical discoveries about preeclampsia. Methods: A detailed search strategy was developed in consultation with a medical librarian. Two bibliographic databases were searched through Ovid: Embase and MEDLINE. We selected case reports or case series published between 2015 and 2020, comprising pregnant persons diagnosed with hypertensive disorders of pregnancy, including preeclampsia. Two reviewers independently screened all publications. One reviewer extracted data from included studies, while another conducted a quality check of extracted data. We developed a codebook to guide our data extraction and outcomes assessment. The quality of each report was determined based on Joanna Briggs Institute (JBI) critical appraisal checklist for case reports and case series. Results: We included 104 case reports and three case series, together comprising 118 pregnancies. A severe presentation or complication of preeclampsia was reported in 81% of pregnancies, and 84% had a positive maternal outcome, free of death or persistent complications. Only 8% of the case reports were deemed to be of high quality, and 53.8% of moderate quality; none of the case series were of high quality. A total of 26 of the 107 publications (24.3%) included a novel clinical discovery as a central theme. Conclusion: Over two-thirds of recent case reports and case series about preeclampsia do not appear to present new knowledge or discoveries about preeclampsia, and most are of low quality.

Anti-vascular endothelial growth factor therapy for age-related macular degeneration: a systematic review and network meta-analysis.

BACKGROUND: The comparative safety and efficacy between anti-vascular endothelial growth factor agents (anti-VEGFs) and between combined therapies for patients with neovascular age-related macular degeneration (nAMD) is unclear. We conducted a systematic review to examine the comparative safety and efficacy anti-VEGFs for adults with nAMD. METHODS: Studies were identified through MEDLINE, EMBASE, and Cochrane CENTRAL (inception to June 3, 2019), grey literature, and scanning reference lists. Two reviewers independently screened citations and full-text articles to identify randomized controlled trials (RCTs), extracted data, and appraised risk of bias. Pairwise random-effects meta-analysis and Bayesian network meta-analysis (NMA) were conducted. The primary outcomes were the proportion of patients experiencing moderate vision gain (≥ 15 letters on the Early Treatment Diabetic Retinopathy Study chart) and the proportion of patients experiencing moderate vision loss (≤ 15 letters). RESULTS: After screening 3647 citations and 485 potentially relevant full-text articles, 92 RCTs with 24,717 patients were included. NMA (34 RCTs, 8809 patients, 12 treatments) showed small differences among anti-VEGFs in improving the proportion of patients with moderate vision gain, with the largest for conbercept versus broluczumab (OR 0.15, 95% CrI: 0.05-0.56), conbercept versus ranibizumab (OR 0.17, 95% CrI: 0.05-0.59), conbercept versus aflibercept (OR 0.19, 95% CrI: 0.06-0.65), and conbercept versus bevacizumab (OR 0.2, 95% CrI: 0.06-0.69). In NMA (36 RCTs, 9081 patients, 13 treatments) for the proportion of patients with moderate vision loss, small differences were observed among anti-VEGFs, with the largest being for conbercept versus aflibercept (OR 0.24, 95% CrI: 0-4.29), conbercept versus brolucizumab (OR 0.24, 95% CrI: 0-4.71), conbercept versus bevacizumab (OR 0.26, 95% CrI: 0-4.65), and conbercept versus ranibizumab (OR 0.27, 95% CrI: 0-4.67). CONCLUSION: The only observed differences were that ranibizumab, bevacizumab, aflibercept, and brolucizumab were statistically superior to conbercept in terms of the proportion of patients with nAMD who experienced moderate vision gain. However, this finding is based on indirect evidence through one small trial comparing conbercept with placebo. This does not account for drug-specific differences when assessing anatomic and functional treatment efficacy in variable dosing regimens. SYSTEMATIC REVIEW REGISTRATION: PROSPERO registration number CRD42015022041.

First-line antihypertensive treatment for severe hypertension in pregnancy: A systematic review and network meta-analysis.

BACKGROUND: Hydralazine, labetalol, and nifedipine are the recommended first-line treatments for severe hypertension in pregnancy. While all three are effective, there is a lack of sufficient evidence regarding their comparative safety and efficacy. OBJECTIVE: To determine the comparative safety and efficacy of the first-line treatment options for severe hypertension in pregnancy. METHODS: A systematic search of Medline, Embase, and Cochrane Central Register of Controlled Trials up to May 31, 2018 was conducted. RCTs in pregnancy comparing a first-line antihypertensive agent to another first-line agent for the treatment of severe hypertension in pregnancy. Screening, data abstraction, and quality assessment were done by two independent reviewers. To estimate relative effects from all available evidence, a Bayesian network meta-analysis with vague priors was conducted. MAIN RESULTS: Of the 1330 publications identified, 17 RCTs comprised of a total of 1591 women met our selection criteria. For successful treatment of severe hypertension, nifedipine was found to be superior to hydralazine (OR 4.13 [95% CrI 1.01-20.75]) but not labetalol (OR 3.43 [95% CrI 0.94-19.95]). This was not associated with an increased risk for caesarean delivery or maternal side effects. There was no significant difference between labetalol and hydralazine. CONCLUSIONS: Given the results of this systematic review and network meta-analysis, maternity care providers should feel comfortable initiating management of severe hypertension in pregnancy using oral nifedipine.

Anti-vascular endothelial growth factor treatment for retinal conditions: a systematic review and meta-analysis.

OBJECTIVES: To evaluate the comparative effectiveness and safety of intravitreal bevacizumab, ranibizumab and aflibercept for patients with choroidal neovascular age-related macular degeneration (cn-AMD), diabetic macular oedema (DMO), macular oedema due to retinal vein occlusion (RVO-MO) and myopic choroidal neovascularisation (m-CNV). DESIGN: Systematic review and random-effects meta-analysis. METHODS: Multiple databases were searched from inception to 17 August 2017. Eligible head-to-head randomised controlled trials (RCTs) comparing the (anti-VEGF) drugs in adult patients aged ≥18 years with the retinal conditions of interest. Two reviewers independently screened studies, extracted data and assessed risk of bias. RESULTS: 19 RCTs involving 7459 patients with cn-AMD (n=12), DMO (n=3), RVO-MO (n=2) and m-CNV (n=2) were included. Vision gain was not significantly different in patients with cn-AMD, DMO, RVO-MO and m-CNV treated with bevacizumab versus ranibizumab. Similarly, vision gain was not significantly different between cn-AMD patients treated with aflibercept versus ranibizumab. Patients with DMO treated with aflibercept experienced significantly higher vision gain at 12 months than patients receiving ranibizumab or bevacizumab; however, this difference was not significant at 24 months. Rates of systemic serious harms were similar across anti-VEGF agents. Posthoc analyses revealed that an as-needed treatment regimen (6-9 injections per year) was associated with a mortality increase of 1.8% (risk ratio: 2.0 [1.2 to 3.5], 2 RCTs, 1795 patients) compared with monthly treatment in cn-AMD patients. CONCLUSIONS: Intravitreal bevacizumab was a reasonable alternative to ranibizumab and aflibercept in patients with cn-AMD, DMO, RVO-MO and m-CNV. The only exception was for patients with DME and low visual acuity (<69 early treatment diabetic retinopathy study [ETDRS] letters), where treatment with aflibercept was associated with significantly higher vision gain (≥15 ETDRS letters) than bevacizumab or ranibizumab at 12 months; but the significant effects were not maintained at 24 months. The choice of anti-VEGF drugs may depend on the specific retinal condition, baseline visual acuity and treatment regimen. PROSPERO REGISTRATION NUMBER: CRD42015022041.

Characteristics of drugs for ultra-rare diseases versus drugs for other rare diseases in HTA submissions made to the CADTH CDR.

BACKGROUND: It has been suggested that ultra-rare diseases should be recognized as distinct from more prevalent rare diseases, but how drugs developed to treat ultra-rare diseases (DURDs) might be distinguished from drugs for 'other' rare diseases (DORDs) is not clear. We compared the characteristics of DURDs to DORDs from a health technology assessment (HTA) perspective in submissions made to the CADTH Common Drug Review. We defined a DURD as a drug used to treat a disease with a prevalence ≤ 1 patient per 100,000 people, a DORD as a drug used to treat a disease with a prevalence > 1 and ≤ 50 patients per 100,000 people. We assessed differences in the level and quantity of evidence supporting each HTA submission, the molecular basis of treatment agents, annual treatment cost per patient, type of reimbursement recommendation made by CADTH, and reasons for negative recommendations. RESULTS: We analyzed 14 DURD and 46 DORD submissions made between 2004 and 2016. Compared to DORDs, DURDs were more likely to be biologic drugs (OR = 6.06, 95%CI 1.25 to 38.58), to have been studied in uncontrolled clinical trials (OR = 23.11, 95%CI 2.23 to 1207.19), and to have a higher annual treatment cost per patient (median difference = CAN$243,787.75, 95%CI CAN$83,396 to CAN$329,050). Also, submissions for DURDs were associated with a less robust evidence base versus DORDs, as DURD submissions were less likely to include data from at least one double-blinded randomized controlled trial (OR = 0.13, 95%CI 0.02 to 0.70) and have smaller patient cohorts in clinical trials (median difference = -108, 95%CI -234 to -50). Furthermore, DURDs are less likely to receive a positive reimbursement recommendation (OR = 0.22, 95%CI 0.05 to 0.91), and low level of evidence was the major contributor for a negative recommendation. CONCLUSIONS: The results suggest that DURDs could be viewed as distinct category from an HTA perspective. Applying the same HTA decision-making framework to DURDs and DORDs might have contributed the higher rate of negative reimbursement recommendations made for DURDs. Recognition of DURDs as a distinct subgroup of DRDs by explicitly defining DURDs based on objective criteria may facilitate the implementation of HTA assessment process that accounts for the issues associated with DURD.

Accelerated Titration of Oxytocin in Nulliparous Women with Labour Dystocia: Results of the ACTION Pilot Randomized Controlled Trial.

OBJECTIVE: The primary objective was to determine the feasibility of a large RCT assessing the effectiveness of an accelerated oxytocin titration (AOT) protocol compared with a standard gradual oxytocin titration (GOT) in reducing the risk of CS in nulliparous women diagnosed with dystocia in the first stage of labour. The secondary objective was to obtain preliminary data on the safety and efficacy of the foregoing AOT protocol. METHODS: This was a multicentre, double-masked, parallel-group pilot RCT. This study was conducted in three Canadian birthing centres. A total of 79 term nulliparous women carrying a singleton pregnancy in spontaneous labour, with a diagnosis of labour dystocia, were randomized to receive either GOT (initial dose 2 mU/min with increments of 2 mU/min) or AOT (initial dose 4 mU/min with increments of 4 mU/min), in a 1:1 ratio. An intention-to-treat analysis was applied. RESULTS: A total of 252 women were screened and approached, 137 (54.4%) consented, and 79 (31.3%) were randomized. Overall protocol adherence was 76 of 79 (96.2%). Of the women randomized, 10 (25.6%) allocated to GOT had a CS compared with six (15.0%) allocated to AOT (Fisher exact test P = 0.27). CONCLUSION: This pilot study demonstrated that a large, multicentre RCT is not only feasible, but also necessary to assess the effectiveness and safety of an AOT protocol for labour augmentation with regard to CS rate and indicators of maternal and perinatal morbidities.

Health technology assessment of drugs for rare diseases: insights, trends, and reasons for negative recommendations from the CADTH common drug review.

BACKGROUND: A shift in biochemical research towards drugs for rare diseases has created new challenges for the pharmaceutical industry, government regulators, health technology assessment agencies, and public and private payers. In this article, we aim to comprehensively review, characterize, identify possible trends, and explore reasons for negative reimbursement recommendations in submissions made to the Common Drug Review (CDR) for drugs for rare diseases (DRD) at the Canadian Agency for Drugs and Technologies in Health (CADTH), a publicly funded pan-Canadian health technology assessment agency. A public database (cadth.ca) was screened to identify DRD submissions to CDR. A diseases prevalence of ≤50 per 100,000 people was considered a rare disease. We calculated descriptive statistics for prevalence, study design, study size, treatment cost, reimbursement recommendation types, and reasons for negative reimbursement recommendations. RESULTS: From 2004 to 2015, 63 of 434 submissions to the CDR were for DRD (range: 1 submission in 2005 to 10 submissions in 2013). Most (74.6%) submissions included at least one double-blind randomized controlled trial (RCT). The average study size was 190 patients (range: 20 to 742). The average annual treatment cost was C$215,631 (range: $9,706 to $940,084). Reimbursement recommendations were positive for 54% of the submissions. Negative reimbursement recommendations were made due to a lack of clinical effectiveness (38.5%), insufficient evidence (30.8%), multiple reasons (23.1%), or lack of cost effectiveness/high cost (7.7%). CONCLUSION: The number of DRD submissions to CDR increased since 2013; from 4 to 5 per year between 2004 and 2012, to 10, 9, and 8 in 2013, 2014, and 2015 respectively. More than half of DRD submissions received positive reimbursement recommendation. Poor quality evidence and/or lack of supportive clinical evidence was at least partly responsible for a negative reimbursement recommendation in all cases. Although the average cost of DRD treatments was high, high cost was a reason for a negative reimbursement recommendation in only two (7.7%) of negative reimbursement recommendations.

Effect of folic acid supplementation during pregnancy on gestational hypertension/preeclampsia: A systematic review and meta-analysis.

OBJECTIVE: To evaluate the effect of folic acid supplementation during pregnancy on the risk of gestational hypertension/preeclampsia. METHODS: A systematic review and meta-analysis were conducted. Medline, Embase, Scopus, and the Web of Science were searched from inception to December 2014. RESULTS: Out of 1224 potentially relevant studies, 13 studies met our inclusion criteria (2 randomized controlled trials (RCTs), 10 cohort studies, and 1 case-control study). The pooled relative risk (RR) and 95% confidence interval (CI) of the two RCTs were 0.62 (0.45-0.87) in the trial arm as compared with the placebo arm. The pooled RR was 0.92 (95% CI: 0.79-1.08) for nine cohort studies with available data on folic acid supplementation in pregnancy and gestational hypertension/preeclampsia. Pooled RR was 0.88 (95% CI: 0.76-1.02) for eight cohort studies with available data on folic acid supplementation and preeclampsia. CONCLUSION: Whether folic acid supplementation in pregnancy can prevent the occurrence of gestational hypertension/preeclampsia remains uncertain.

Factors Associated With Increased Rates of Caesarean Section in Women of Advanced Maternal Age.

OBJECTIVE: To compare rates of Caesarean section between mothers of advanced age (35 to 40, and over 40 years) and those aged 20 to 34, using the Robson classification system to examine additional maternal factors. METHODS: A total of 134 088 hospital deliveries in Ontario between April 1, 2011, and March 31, 2012, were grouped into Robson's 10 mutually exclusive and totally inclusive classification categories. Records from the three Robson groups that made the greatest contribution to the overall CS rate were stratified by maternal age, health condition, obstetrical complication, assisted reproductive technology usage, smoking during pregnancy, and socioeconomic status. RESULTS: Rates of CS increased with advancing maternal age; in women aged 20 to 34, 35 to 40, and over 40, the rates were 26.2%, 35.9%, and 43.1%, respectively. The top three Robson groups by contribution to CS rates involved women who had one or more of the following factors: previous Caesarean section, primiparity, conception by means of assisted reproductive technology, chronic hypertension, gestational diabetes, diabetes mellitus, preeclampsia, placenta previa, placental abruption, or large for gestational age infants. The prevalence of these factors increased with advancing maternal age, yet mothers aged ≥ 35 with one or more health conditions or obstetrical complications had higher CS rates than mothers aged 20 to 34 with the same condition(s) or complication(s). CONCLUSION: Health conditions and obstetrical complications alone in older women do not account for increased rates of CS. The preferences of the individual care provider and the mother on CS rates may play a key role and require further investigation.

Objectif : Comparer les taux de césarienne des mères d'âge avancé (de 35 à 40 ans et de plus de 40 ans) à ceux des mères âgées de 20 à 34 ans, en utilisant le système de classification de Robson en vue d'examiner des facteurs maternels additionnels. Méthodes : Au total, 134 088 accouchements s'étant déroulés en milieu hospitalier en Ontario entre le 1er avril 2011 et le 31 mars 2012 ont été groupés en fonction des 10 catégories mutuellement exclusives et totalement inclusives de Robson. Les dossiers des trois groupes Robson ayant le plus contribué au taux global de césarienne ont été stratifiés en fonction de l'âge maternel, de l'état de santé, des complications obstétricales, du recours à des techniques de procréation assistée, du tabagisme pendant la grossesse et du statut socioéconomique. Résultats : Les taux de césarienne étaient proportionnels à l'âge maternel : chez les femmes de 20 à 34 ans, de 35 à 40 ans et de plus de 40 ans, les taux ont été de 26,2 %, de 35,9 % et de 43,1 %, respectivement. Les trois groupes Robson ayant le plus contribué au taux global de césarienne étaient composés de femmes qui présentaient un ou plusieurs des facteurs suivants : antécédents de césarienne, primiparité, conception au moyen de techniques de procréation assistée, hypertension chronique, diabète gestationnel, diabète sucré, prééclampsie, placenta praevia, décollement placentaire ou hypertrophie fœtale. Bien que la prévalence de ces facteurs ait été proportionnelle à l'âge maternel, les mères âgées de 35 ans ou plus qui comptaient un ou plusieurs troubles de santé (ou complications obstétricales) présentaient des taux de césarienne supérieurs à ceux des mères âgées de 20 à 34 ans qui comptaient le ou les mêmes troubles (ou complications). Conclusion : Les taux accrus de césarienne chez les femmes plus âgées ne peuvent être attribués qu'à la seule présence de troubles de santé et de complications obstétricales. Les préférences des fournisseurs de soins et des mères en matière d'accouchement pourraient jouer un rôle clé en ce qui concerne les taux de césarienne, ce qui nécessite la tenue d'études plus approfondies.