Application of machine learning methods to bridge the gap between non-interventional studies and randomized controlled trials in ophthalmic patients with neovascular age-related macular degeneration

PURPOSE

The effectiveness of intravitreal anti-vascular endothelial growth factor agents is usually lower in real world settings compared with randomized clinical trials (RCTs), often limiting the use of real-world evidence (RWE) in regulatory and healthcare decisions. The current analysis aimed to develop and validate an algorithm to explain the difference in outcomes between RWE studies and RCTs in patients with neovascular age-related macular degeneration.

METHODS

The algorithm was developed using ranibizumab real world data (RWD) from the US and validated on Australian and UK RWD. A decision model was developed using machine learning principles, in which the model learns how to partition the most influential factors (out of 59 variables) so that they maximally relate to the change in visual acuity (VA) over 12 months.

RESULTS

The algorithm identified baseline VA <73 Early Treatment Diabetic Retinopathy Study letters, presence of baseline subretinal fluid, and administration of three loading doses by Day 90 from drug initiation as the characteristics with the greatest impact on VA at month 12. When applying the different criteria, RWE outcomes became similar to those obtained in known RCTs.

CONCLUSION

Machine learning techniques can be used to classify real world cohorts and identify subsets of patients who benefit to the same extent as that reported in RCTs. This methodology may support the translation of clinical trial findings to treatment performance in the clinical practice setting.

Time to Revisit a Voluntary FDA Comparative Effectiveness Pathway

Given the renewed policy focus on drug pricing and pharmaceutical innovation, this article examines the historical backdrop of efforts to integrate comparative effectiveness research into the FDA drug review process. Noting previous policy efforts over a decade ago, we characterize industry challenges and suggest a path forward.

Demonstrating that Real World Evidence Is Fit-For-Purpose to Support Labeling: Parallels to Patient Reported Outcomes in the Pursuit of Labeling Claims

INTRODUCTION

In December 2021, U.S. Food & Drug Administration (FDA) will issue guidance on the use of real-world evidence (RWE) to support new indications or expanded product labeling. While difficult to foresee what FDA will require, learnings can be gleaned from previous paradigm shifts at FDA, such as for patient reported outcomes (PROs) in 2006-2009.

METHODS

We contrast published requirements for justifying PROs as fit-for-purpose for a specific labeling claim with a potential approach to justify RWE as fit-for-purpose to support expanded labeling or a new indication.

RESULTS

PRO labeling claims require a PRO Evidence Dossier that includes: specific wording of claim, clinical trial hypothesis structure and endpoint model, and justification that the PRO is relevant and meaningful to patients in the target population (content validity) with adequate psychometric properties. FDA's 2018 RWE Framework outlined critical considerations for using RWE to support regulatory decisions, including data quality, relevancy, provenance, and transparency. Strong parallels exist between the evidence required to justify that PROs are fit-for-purpose to support specific labeling claims and evidence to justify RWE as fit-for-purpose for specific research questions and labeling. Early discussion with FDA is encouraged.

CONCLUSION

Drawing on parallels with use of PROs in labeling, RWE for regulatory purposes should be evaluated within the context of specific labeling or indication, specific study design and analysis plans, and the data attributes of data source. Sponsors seeking a new indication or labeling expansion based on RWE should justify that a specific data source and specific study design are fit-for-purpose.

Real-world evidence on Kovaltry (81-8973) in children with moderate or severe hemophilia A in Europe: a nested cohort analysis

Background

Untreated hemophilia A patients may experience recurrent bleeding events leading to debilitating joint damages. While RCT and pharmacokinetic data support the value of Kovaltry [an unmodified full-length recombinant factor VIII (FVIII) product], real world evidence in children is lacking. This report describes a descriptive and multivariate analysis of the effectiveness of Kovaltry in children with hemophilia A in the real-world setting, using data from medical chart abstraction and cross-sectional surveys of physicians, patients, and caregivers.

Results

Male patients aged < 18 years with moderate or severe hemophilia A, residing in five European countries and treated with FVIII were studied. The co-primary endpoints were the annualized bleeding rate (ABR) and the annual FVIII utilization rate. Twenty nine patients treated with Kovaltry were included, of whom 93% had severe disease and 75% were on continuous prophylactic treatment. The mean ABR was 2.66 ± 2.06, with rates decreasing with age. The children received on average 2.45 infusions per week, consistent across age groups (median 3; range 1–3). There were no reports of inhibitor development or adverse events in the study (AEs), and all patients were satisfied or very satisfied with the treatment. An exploratory multivariate analysis suggests no significant difference in ABR or units utilized between Kovaltry and some extended half life products in children with severe hemophilia A, though characteristics of these patient cohorts were markedly different.

Conclusion

This analysis demonstrates the effectiveness and safety of Kovaltry in a pan-European pediatric population with severe hemophilia A.

Transforming vaccine development

The urgency to develop vaccines against Covid-19 is putting pressure on the long and expensive development timelines that are normally required for development of lifesaving vaccines. There is a unique opportunity to take advantage of new technologies, the smart and flexible design of clinical trials, and evolving regulatory science to speed up vaccine development against Covid-19 and transform vaccine development altogether.

Cohort profile: King's Health Partners bladder cancer biobank

BACKGROUND

Bladder cancer (BC) is the 9th most common cancer worldwide, but little progress has been made in improving patient outcomes over the last 25 years. The King's Health Partners (KHP) BC biobank was established to study unanswered, clinically relevant BC research questions. Donors are recruited from the Urology or Oncology departments of Guy's Hospital (UK) and can be approached for consent at any point during their treatment pathway. At present, patients with bladder cancer are approached to provide their consent to provide blood, urine and bladder tissue. They also give access to medical records and linkage of relevant clinical and pathological data across the course of their disease. Between June 2017 and June 2019, 531 out of 997 BC patients (53.3%) gave consent to donate samples and data to the Biobank. During this period, the Biobank collected fresh frozen tumour samples from 90/178 surgical procedures (of which 73 were biopsies) and had access to fixed, paraffin embedded samples from all patients who gave consent. Blood and urine samples have been collected from 38 patients, all of which were processed into component derivatives within 1 to 2 h of collection. This equates to 193 peripheral blood mononuclear cell vials; 238 plasma vials, 224 serum vials, 414 urine supernatant vials and 104 urine cell pellets. This biobank population is demographically and clinically representative of the KHP catchment area.

CONCLUSION

The King's Health Partners BC Biobank has assembled a rich data and tissue repository which is clinically and demographically representative of the local South East London BC population, making it a valuable resource for future BC research.

Clinical Characteristics, Treatments, and Concurrent Mutations in Non-Small Cell Lung Cancer Patients With NF1 Mutations

OBJECTIVES

Metastatic non-small cell lung cancer (mNSCLC) is characterized by complex genomic alterations. NF1 mutations may confer distinct clinical characteristics within NSCLC, and real-world evidence on concurrent mutations, treatment patterns, and health outcomes is lacking.

MATERIALS AND METHODS

This retrospective study was performed in patients with mNSCLC treated in the Flatiron Health network who underwent the FoundationOne tumor-sequencing. Anticancer therapies, concurrent mutations, real-world progression-free survival (rwPFS), and overall survival (OS) were assessed.

RESULTS

Of the 1663 patients, 103 patients were identified with NF1 mutation. Concurrent mutations with Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (16.5%) and epidermal growth factor receptor fusion (6.8%) were the most frequent. In patients with NF1 mutation only (n = 57), 42% were women, 86% patients had smoking history, and 70% had non-squamous cell carcinoma type. Most (51%) of the patients with NF1 mutations received a single line of therapy versus other mutations and the overall treated population (44%). Platinum-based chemotherapy was the predominant first-line therapy, with programmed cell death-1/programmed cell death-ligand-1 inhibitors as subsequent lines of therapy. The NF1 mutation only group had numerically the shortest median rwPFS (82 days) than other mutation groups. Median OS for the NF1 mutation group in first, second, and third lines of therapy was 321, 498, and 210 days, respectively.

CONCLUSIONS

NF1 mutations confer distinct clinical characteristics in patients with mNSCLC. These patients may have different trajectories for progression and survival than seen for other mutations, experience less systemic therapy after first-line therapy, and may have shorter survival.

Socioeconomic factors associated with hypoglycaemia in patients starting second-line glucose-lowering therapy: The DISCOVER study

AIMS

Using data from DISCOVER (NCT02322762; NCT02226822), a 3-year, global, observational study programme of patients with type 2 diabetes initiating second-line glucose-lowering therapy, we assessed socioeconomic factors associated with hypoglycaemic events and fear of hypoglycaemia.

METHODS

Data were collected at baseline (second-line therapy initiation) and 6, 12 and 24 months. Factors associated with experiencing a hypoglycaemic event at baseline or during follow-up were determined using a hierarchical logistic regression model and an interval-censored survival analysis, respectively. Fear of hypoglycaemia was assessed using the hypoglycaemia fear survey-II (HFS-II).

RESULTS

The overall proportion of patients reporting hypoglycaemic events during follow-up was 7.3%; this was higher in middle-income countries than in high-income countries (8.4% vs 5.8%, p < 0.001). Factors associated with an increased risk of hypoglycaemia during follow-up included living in a country with a low gross national income, use of glucose-monitoring equipment and second-line treatment with insulin, meglitinides or sulphonylureas (versus metformin). Experiencing hypoglycaemia was associated with increased HFS-II worry and overall scores.

CONCLUSIONS

Our results highlight the global inequity in the treatment of type 2 diabetes. Increased risk of hypoglycaemia in middle-income countries may be explained by limited treatment options and may be underestimated because of limited access to glucose-monitoring equipment.

Multicentric, prospective observational data show sperm capacitation predicts male fertility, and cohort comparison reveals a high prevalence of impaired capacitation in men questioning their fertility

RESEARCH QUESTIONS

Can a previously defined relationship between sperm capacitation and the probability of a man generating pregnancy within three cycles, prospectively predict male fertility in diverse clinical settings? A second study asked, what is the prevalence of impaired sperm fertilizing ability in men questioning their fertility (MQF), and does this relate to traditional semen analysis metrics?

DESIGN

In the multicentric, prospective observational study, data (n = 128; six clinics) were analysed to test a published relationship between the percentage of fertilization-competent, capacitated spermatozoa (Cap-Score) and probability of generating pregnancy (PGP) within three cycles of intrauterine insemination. Logistic regression of total pregnancy outcomes (n = 252) assessed fit. In the cohort comparison, Cap-Scores of MQF (n = 2155; 22 clinics) were compared with those of 76 fertile men.

RESULTS

New outcomes (n = 128) were rank-ordered by Cap-Score and divided into quintiles (25-26 per group); chi-squared testing revealed no difference between predicted and observed pregnancies (P = 0.809). Total outcomes (n = 252; 128 new + 124 previous) were pooled and the model recalculated, yielding an improved fit (P < 0.001). Applying the Akaike information criterion found that the optimal model used Cap-Score alone. Cap-Scores were performed on 2155 men (with semen analysis data available for 1948). To compare fertilizing ability, men were binned by PGP (≤19%, 20-29%, 30-39%, 40-49%, 50-59%, ≥60%). Distributions of PGP and the corresponding Cap-Scores were significantly lower in MQF versus fertile men (P < 0.001). Notably, 64% of MQF with normal volume, concentration and motility (757/1183) had PGP of 39% or less (Cap-Scores ≤31), versus 25% of fertile men.

CONCLUSIONS

Sperm capacitation prospectively predicted male fertility. Impaired capacitation affects many MQF with normal semen analysis results, informing diagnosis versus idiopathic infertility.

How an electronic health record became a real-world research resource: comparison between London's Whole Systems Integrated Care database and the Clinical Practice Research Datalink

BACKGROUND

In the UK, several initiatives have resulted in the creation of local data warehouses of electronic patient records. Originally developed for commissioning and direct patient care, they are potentially useful for research, but little is known about them outside their home area. We describe one such local warehouse, the Whole Systems Integrated Care (WSIC) database in NW London, and its potential for research as the "Discover" platform. We compare Discover with the Clinical Practice Research Datalink (CPRD), a popular UK research database also based on linked primary care records.

METHODS

We describe the key features of the Discover database, including scope, architecture and governance; descriptive analyses compare the population demographics and chronic disease prevalences with those in CPRD.

RESULTS

As of June 2019, Discover held records for a total of 2.3 million currently registered patients, or 95% of the NW London population; CPRD held records for over 11 million. The Discover population matches the overall age-sex distribution of the UK and CPRD but is more ethnically diverse. Most Discover chronic disease prevalences were comparable to the national rates. Unlike CPRD, Discover has identifiable care organisations and postcodes, allowing mapping and linkage to healthcare provider variables such as staffing, and includes contacts with social, community and mental health care. Discover also includes a consent-to-contact register of over 3000 volunteers to date for prospective studies.

CONCLUSIONS

Like CPRD, Discover has been a number of years in the making, is a valuable research tool, and can serve as a model for other areas developing similar data warehouses.

Guy's cancer cohort - real world evidence for cancer pathways

Background

The burden of disease due to cancer remains substantial. Since the value of real-world evidence has also been recognised by regulatory agencies, we established a Research Ethics Committee (REC) approved research database for cancer patients (Reference: 18/NW/0297).

Construction and content

Guy’s Cancer Cohort introduces the concept of opt-out consent processes for research in a subset of oncology patients diagnosed and treated at a large NHS Trust in the UK. From April 2016 until March 2017, 1388 eligible patients visited Guy’s and St Thomas’ NHS Foundation Trust (GSTT) for breast cancer management. For urological cancers this number was 1757 and for lung cancer 677. The Cohort consists of a large repository of routinely collected clinical data recorded both retrospectively and prospectively. The database contains detailed clinical information collected at various timepoints across the treatment pathway inclusive of diagnostic data, and data on disease progression, recurrence and survival.

Conclusions

Guy’s Cancer Cohort provides a valuable infrastructure to answer a wide variety of research questions of a clinical, mechanistic, and supportive care nature. Clinical research using this database will result in improved patient safety and experience. Guy’s Cancer Cohort promotes collaborative research and will accept applications for the release of anonymised datasets for research purposes.

A roadmap to using historical controls in clinical trials - by Drug Information Association Adaptive Design Scientific Working Group (DIA-ADSWG)

Historical controls (HCs) can be used for model parameter estimation at the study design phase, adaptation within a study, or supplementation or replacement of a control arm. Currently on the latter, there is no practical roadmap from design to analysis of a clinical trial to address selection and inclusion of HCs, while maintaining scientific validity. This paper provides a comprehensive roadmap for planning, conducting, analyzing and reporting of studies using HCs, mainly when a randomized clinical trial is not possible. We review recent applications of HC in clinical trials, in which either predominantly a large treatment effect overcame concerns about bias, or the trial targeted a life-threatening disease with no treatment options. In contrast, we address how the evidentiary standard of a trial can be strengthened with optimized study designs and analysis strategies, emphasizing rare and pediatric indications. We highlight the importance of simulation and sensitivity analyses for estimating the range of uncertainties in the estimation of treatment effect when traditional randomization is not possible. Overall, the paper provides a roadmap for using HCs.

Regulatory Science, and How Device Regulation Will Shape Our Future

Pediatric medical device approvals lag behind adult approvals. Historically, medical devices have rarely been designed specifically for children, but use in children has most often borrowed from adult or general use applications. While a variety of social, economic, and clinical factors have contributed to this phenomenon, the regulatory process remains a fundamental aspect of pediatric device development and commercialization. FDA's Center for Devices and Radiological Health (CDRH) has established programmatic and technological areas of advancement to support innovation that serves the public health needs of children and special populations. We highlight four regulatory areas that have the potential to shape the future of pediatric cardiology: the CDRH Early Feasibility Study Program, advancements in 3D printing or additive manufacturing, computational modeling and simulation, and the use of real-world evidence for regulatory applications. These programs have the potential to impact all stages of device development, from early conception, design, and prototyping to clinical evidence generation, regulatory review, and finally commercialization. The success of these programs relies on a collaborative community of stakeholders, including government, regulators, device manufacturers, patients, payers, and the academic and professional community societies.

ADVANCE database characterisation and fit for purpose assessment for multi-country studies on the coverage, benefits and risks of pertussis vaccinations

INTRODUCTION

The public-private ADVANCE consortium (Accelerated development of vaccine benefit-risk collaboration in Europe) aimed to assess if electronic healthcare databases can provide fit-for purpose data for collaborative, distributed studies and monitoring of vaccine coverage, benefits and risks of vaccines.

OBJECTIVE

To evaluate if European healthcare databases can be used to estimate vaccine coverage, benefit and/or risk using pertussis-containing vaccines as an example.

METHODS

Characterisation was conducted using open-source Java-based (Jerboa) software and R scripts. We obtained: (i) The general characteristics of the database and data source (meta-data) and (ii) a detailed description of the database population (size, representatively of age/sex of national population, rounding of birth dates, delay between birth and database entry), vaccinations (number of vaccine doses, recording of doses, pattern of doses by age and coverage) and events of interest (diagnosis codes, incidence rates). A total of nine databases (primary care, regional/national record linkage) provided data on events (pertussis, pneumonia, death, fever, convulsions, injection site reactions, hypotonic hypo-responsive episode, persistent crying) and vaccines (acellular pertussis and whole cell pertussis) related to the pertussis proof of concept studies.

RESULTS

The databases contained data for a total population of 44 million individuals. Seven databases had recorded doses of vaccines. The pertussis coverage estimates were similar to those reported by the World Health Organisation (WHO). Incidence rates of events were comparable in magnitude and age-distribution between databases with the same characteristics. Several conditions (persistent crying and somnolence) were not captured by the databases for which outcomes were restricted to hospital discharge diagnoses.

CONCLUSION

The database characterisation programs and workflows allowed for an efficient, transparent and standardised description and verification of electronic healthcare databases which may participate in pertussis vaccine coverage, benefit and risk studies. This approach is ready to be used for other vaccines/events to create readiness for participation in other vaccine related studies.

Survival benefit of skip metastases in surgically resected N2 non-small cell lung cancer: A multicenter observational study of a large cohort of the Chinese patients

INTRODUCTION

N2 non-small cell lung cancer (NSCLC) without N1 involvement, also known as skip metastases (pN0N2), has been suggested as a subgroup of heterogeneous N2 disease with better survival. This real-world observational study aimed to elucidate the prognostic impacts of skip N2 metastases using a large number of pathologic N2 NSCLC from 10 participating centers in China.

MATERIALS AND METHODS

Medical records of pN2 NSCLC patients after surgical resection at 10 thoracic surgery centers between January 2014 and September 2017 were retrospectively reviewed based on the LinkDoc database. Clinical data on patient demographics, tumor characteristics, treatments and clinical outcomes were collected. Overall survival of patients with and without skip metastases was evaluated and compared by Kaplan-Meier method and Log-rank test. Cox proportional hazard model was established to identify potential prognostic predictors. Subgroup analysis was carried out to further explore the prognostic significance of skip metastases.

RESULTS

Among 2653 surgically resected N2 patients, 881 (33.2%) had skip metastases. Patients with skip N2 had a significant better overall survival (P = 0.0019). Multivariate COX regression analysis showed borderline significance of skip metastases (HR = 0.81, 95%CI: 0.645-1.017, P = 0.0698) after adjustment for other covariates. Other independent prognostic predictors included smoking history, tumor location, stage and N2 station involved (P < 0.05). Subgroup analysis demonstrated significant survival benefits of skip N2 in most subpopulations.

CONCLUSIONS

This study suggested a prognostic benefit of skip N2 metastases in real world practice. Further subdivision of N2 disease is warranted for better patient management and prognostic prediction (NCT03429192).

Indirect costs associated with ulcerative colitis: a systematic literature review of real-world data

Background

The economic burden of ulcerative colitis (UC), specifically related to indirect costs, is not extensively documented. Understanding and quantifying it is required by health care decision makers.

Aim

To assess the impact of indirect costs of UC in observation studies.

Method

A systematic literature search was conducted in MEDLINE®, Embase® and Cochrane Library to capture all relevant publications reporting outcomes on absenteeism, presenteeism and productivity losses in moderate to severe UC. Eligibility criteria for inclusion into the review were established using a predefined PICOS scheme. All costs were adjusted to 2017 currency values (USD dollars, $).

Results

In total, 18 studies reporting data on indirect costs were included in the analysis. Absenteeism costs were classified into three categories: sick leave, short-term and long-term disability. Most of the studies captured absenteeism costs related specifically to sick leave, which was experienced on average by 10 to 24% patients with UC. Only three studies captured presenteeism costs, as these are difficult to measure, however costs ranged from 1602 $ to 2947 $ per patient year. The proportion of indirect costs accounted for 35% of total UC costs (Total UC costs were defined as the sum of healthcare costs, productivity costs and out-of-pocket costs).

Discussion

A limited number of studies were identified describing the indirect costs in patients with moderate to severe UC. Insufficient data on different components of costs allowed a limited analysis on the impact of indirect costs in patients with UC. Further studies are needed to gain an understanding of the influence of UC on patients’ functional abilities.

Hybrid clinical trials to generate real-world evidence: design considerations from a sponsor's perspective

Randomized controlled trials have traditionally been the gold standard for evaluating efficacy and safety of medical products and for regulatory decision-making. With the advancement of information technologies, vast amounts of data pertinent to patient health status and health care delivery are becoming available from a variety of real-world sources, including electronic health records, medical claims, patient registries, and patient-generated data. In 2016, the United States Congress passed the 21st Century Cures Act, mandating the U.S. FDA to establish a program to evaluate the potential use of real-world evidence (RWE) for regulatory purposes. In 2018, the FDA published the framework on its RWE program. One particular study type identified in the framework is the hybrid design - integration of a traditional randomized controlled trial with pragmatic design aspects to collect real-world data on patients. This design preserves the benefit of randomization, provides real-world outcome data while potentially accelerating product development and lowering the cost of data collection and patient follow-up. Here we focus on design considerations for hybrid trials to support regulatory decisions and provide a sponsor's perspective. While applicable to all medical products, we emphasize vaccine development where such hybrid designs are particularly useful given the low incidence rate of some vaccine-preventable clinical outcomes. We propose program strategies on how such hybrid designs may be integrated into a clinical development plan, illustrated by three examples. Major challenges are discussed and recommendations provided. Given the promise of hybrid designs and the challenges in implementation, we encourage proactive discussion with health authorities.

Pharmacovigilance strategy: opportunities for cross-national learning

The Israeli Ministry of Health has set up the foundations of a National Pharmacovigilance System. The next step is to adopt the best of the international ideas, trends and approaches which are shaping the future of pharmacovigilance. Specifically: 1) The risk management approach requires proactively preventing or minimizing risks, starting in early clinical development and extending all along the lifecycle of a pharmaceutical. 2) Drug safety is a multidisciplinary discipline where all stakeholders should be involved. 3) Clinical trials provide an ideal safety profile limited to the restrictive conditions of the trial. Only real-world data, from the post marketing period, will reveal the real risk/benefit balance for the use of a pharmaceutical in regular clinical care. 4) Artificial intelligence is needed to analyze the large amount of data collected through the post-marketing studies, electronic medical records and the internet. Many AI tools have been developed to support better use of pharmaceuticals. 5) Quality-oriented, thorough inspections and audits are critical for achieving patient safety. 6) Patients should be recognized as active players in their treatment who can, and should, have access to safety information through the major agencies’ websites.

Israel can benefit from several of its key assets to reach a higher level of pharmacovigilance: 1) Israel’s four HMOs are organized in a way that allows them to have quick and efficient dialogue with healthcare professionals and with patients. Moreover, a new project named, Big Data in Health, will pool the epidemiologic databases of the HMOs, providing precious information for understanding risk factors, detecting alerts, and developing personalized medicine. 2) Formal risk management activities have long been part of the culture of hospitals and should be applied increasingly to ensuring drug safety.

Israel has the organizational, scientific, technological and cultural resources needed to quickly overcome the challenges and go beyond its current state to build a unique pharmacovigilance system which could serve as an example for other countries.

Real-world evidence was feasible for estimating effectiveness of chemotherapy in breast cancer: a cohort study

OBJECTIVE

Evidence-based guidelines recommend adjuvant chemotherapy in early stage breast cancer whenever treatment benefit is considered sufficient to outweigh the associated risks. However, many groups of patients were either excluded from or underrepresented in the clinical trials that form the evidence base for this recommendation. This study aims to determine whether using administrative health care data-real world data-and econometric methods for causal analysis to provide "real world evidence" (RWE) are feasible methods for addressing this gap.

METHODS

Cases of primary breast cancer in women from 2001 to 2015 were extracted from the Scottish cancer registry (SMR06) and linked to other routine health records (inpatient and outpatient visits). Four methods were used to estimate the effect of adjuvant chemotherapy on disease-specific and overall mortality: (1) regression with adjustment for covariates, (2) propensity score matching, (3) instrumental variables analysis, and (4) regression discontinuity design. Hazard ratios for breast cancer mortality and all-cause mortality were compared to those from a meta-analysis of randomized trials.

RESULTS

A total of 39,805 cases were included in the analyses. Regression adjustment, propensity score matching, and instrumental variables were feasible, whereas regression discontinuity was not. Effectiveness estimates were similar between RWE and randomized trials for breast cancer mortality but not for all-cause mortality.

CONCLUSIONS

RWE methods are a feasible means to generate estimates of effectiveness of adjuvant chemotherapy in early stage breast cancer. However, such estimates must be interpreted in the context of the available randomized evidence and the potential biases of the observational methods.

Using real world data to assess cardiovascular outcomes of two antidiabetic treatment classes

AIM

To evaluate the effect on cardiovascular outcomes of sodium-glucose co-transporter-2 (SGLT2) inhibitors in a real world setting by analyzing electronic medical records.

METHODS

We used TriNetX, a global federated research network providing statistics on electronic health records (EHR). The analytics subset contained EHR from approximately 38 Million patients in 35 Health Care Organizations in the United States. The records of 46,909 patients who had taken SGLT2 inhibitors were compared to 189,120 patients with dipeptidyl peptidase (DPP) 4 inhibitors. We identified five potential confounding factors and built respective strata: elderly, hypertension, chronic kidney disease (CKD), and co-medication with either insulin or metformin. Cardiovascular events were counted as stroke (ICD10 code: I63) or myocardial infarction (ICD10: I21) occurring within three years after the first instance of the respective medication in the patients' records.

RESULTS

Of the 46909 patients with SGLT2 inhibitors in their EHR, 1667 patients (3.6%) had an ICD code for stroke or for myocardial infarction within the first three years after the first instance of the medication. In the control group, there were 10680 events of 189120 patients (5.6%), which represents a risk ratio of 0.63 (95%CI: 0.60-0.66). The overall incidence of stroke or myocardial infarction in the strata with a potential confounding risk factor reached from 4.9% in patients taking metformin to 12.5% in the stratum with the highest risk (concomitant CKD). In all strata, the difference in risk of experiencing a cardiovascular event was similarly in favor of SGLT2 vs control, with Risk Ratio ranging from 0.62 to 0.81.

CONCLUSION

Real world data replicated the results from randomized clinical trials, confirmed the cardiovascular advantages of SGLT2 inhibitors, and showed its applicability to the US population.

Liraglutide and cardiovascular outcomes in a real world type 2 diabetes cohort

In the last years, due to new regulatory guidelines requiring a stringent documentation of cardiovascular (CV) safety of novel drugs for type 2 diabetes, cardiovascular outcomes safety trials (CVOTs) are requested. CVOTs increase the knowledge about the safety profile of the new drugs, but they have intrinsic limits that make difficult their transferability to clinical practice. For this reason, real world evidence is considered an important complement to experimental data. Among the glucagon-like peptide-1 receptor agonists, liraglutide in the LEADER CVOT demonstrated superiority in reducing the risk of major CV events (MACEs) vs. placebo. We conducted an observational, retrospective, longitudinal study based on 307 patients with T2DM treated with liraglutide under routine clinical practice conditions. Real world impact of liraglutide on metabolic control, CV risk factors, hypoglycemia and CV events was assessed. Improvements during 36 months were found in HbA_1c (-1.0%; p < 0.0001), fasting blood glucose (-17.6 mg/dL; p < 0.0001), body weight (-3.2 kg; p < 0.0001), waist circumference (-1.45 cm; p = 0.004), systolic blood pressure (-10.41 mmHg; p < 0.0001), diastolic blood pressure (-3.69 mmHg; p < 0.0001), total cholesterol (-7.96 mg/dL; p =0.008) and triglycerides (-20.60 mg/dl; p = 0.01). No severe hypoglycemia occurred. Incidence of MACEs in this cohort was lower than in the LEADER study (2.59 vs. 3.4 events per 100 person-years), confirming CV safety of liraglutide even in the real world. On the other hand, a higher incidence of CV event in patients with established CV disease was documented (8.1 events per 100 person-years), in spite of the use of liraglutide. In conclusion, 36-month durability and CV safety of liraglutide were documented in a real world cohort of T2DM patients, with sustained benefits on a large array of CV risk factors.

Xafran, a drug utilization study of rivaroxaban in stroke prevention in atrial fibrillation in France using a claim database

BACKGROUND

Little is known about the utilization of rivaroxaban in real life treatment settings in France.

AIMS

Objectives of this study were to describe the conditions of use and treatment persistence in patients with atrial fibrillation and treated with rivaroxaban for stroke prevention (SP-AF).

METHODS

A cohort study was performed using a representative sample of the French nationwide database. All adults who initiated rivaroxaban for SP-AF between 01/08/2012 and 31/12/2014 were included and followed for one year. Inappropriate use of rivaroxaban's was defined as use inconsistent with the summary of product characteristics.

RESULTS

In this study, 1278 patients were included, 687 (53.8%) were men and the mean age was 73.4years; 123 patients (9.6%) had a stroke and 78 (6.1%) a major bleeding event in the three years before rivaroxaban initiation. At treatment initiation 236 (18.5%) had chronic congestive heart failure, 991 (77.5%) hypertension, 247 (19.3%) diabetes and 9 (0.7%) HIV, hepatitis B or C infection. No anticoagulant had been administered in the six previous months for 777 patients (60.8%); 160 patients (12.5%) had an inappropriate use of rivaroxaban in SP-AF. At 6 and 12 months after the first delivery with rivaroxaban, 62.8%, and 51.7% (68.5% and 60.5% in sensitivity analyses) of the patients were still treated with rivaroxaban. The proportion of patients with a continuous medication availability above 80% was 96.1%.

CONCLUSION

The characteristics of patients in this study are similar to patients treated with this drug in other observational studies. Adherence and persistence with rivaroxaban can be considered good.

[Pro re nata anti-VEGF treatment results for neovascular age-related macular degeneration in routine clinical treatment: comparison of single with triple injections]

BACKGROUND

Different injection regimens from continuous to pro re nata (PRN) have been proposed for treatment of neovascular age-related macular degeneration (nAMD). So far the PRN single injection on reactivation regimen has not been compared to the PRN triple injection on reactivation regimen (IVAN scheme).

OBJECTIVE

Comparison of the two nAMD PRN injection regimens with single and triple injections on reactivation in a real-world setting in a retrospective case series in two German treatment centers.

MATERIAL AND METHODS

Naïve nAMD patients, who started treatment according to either the single or triple injection regimen were included. Endpoints were best corrected visual acuity (LogMAR), central retinal thickness on optical coherence tomography (μm) and number of injections, all at 3, 6, 12, 18 and 24 months after treatment initiation.

RESULTS

A total of 146 patients with single injection and 148 patients with triple injection regimens were included. There were no significant differences between the two treatment regimens in best corrected visual acuity (single vs. triple injection scheme: 0.50 ± 0.42 vs. 0.56 ± 0.42, p = 0.14), central retinal thickness (303 ± 76.2 vs. 306 ± 110, p = 0.79) and number of injections (13 ± 4.4 vs. 12 ± 5.4, p = 0.31). This was the case for all analyzed time points.

CONCLUSION

There were no significant functional or morphological differences between the two PRN injection regimens with single and triple injections on reactivation after 24 months. For evaluation of long-term therapy results further studies are warranted.

What have we learned from real-world NOAC studies in venous thromboembolism treatment?

Venous thromboembolism (VTE) remains a substantial clinical and health-economic burden worldwide and effective anticoagulant treatment is necessary immediately after VTE is suspected to reduce short- and long-term VTE related morbidity and mortality. For decades, low molecular weight heparin (LMWH), fondaparinux and Vitamin K antagonists (VKAs) have been the standard of anticoagulant therapy for VTE patients but these treatment options had clinically relevant drawbacks and limitations. The introduction of non-VKA oral anticoagulants (NOACs) that specifically inhibit either thrombin or factor Xa have resolved many of these drawbacks because these new compounds exhibit a rapid onset and offset of action, fewer food and drug interactions and a predictable anticoagulant effect. All NOACs have successfully completed their respective phase-III trial programs consisting of many large randomized controlled trials, leading to approval for acute VTE treatment around the world. Nevertheless, their introduction into daily care practice is challenging and a careful evaluation of the effectiveness and safety of NOACs in less selected cohorts outside carefully monitored clinical trials is essential. This review introduces the different types of real-world evidence (RWE) and explores the available data for VTE treatment with NOACs, based on a literature search using the key words "venous thromboembolism" or "VTE" in combination with "NOAC", "DOAC", "apixaban", "dabigatran", "edoxaban" and "rivaroxaban" on June 30; 2017, followed by data extraction from studies that reported real-world outcome data for VTE treatment with NOACs, although available evidence is almost exclusively limited to rivaroxaban.

Real world evidence (RWE) - a disruptive innovation or the quiet evolution of medical evidence generation?

Stakeholders in healthcare are increasingly turning to real world evidence (RWE) to inform their decisions, alongside evidence from randomized controlled trials. RWE is generated by analysing data gathered from routine clinical practice, and can be used across the product lifecycle, providing insights into areas including disease epidemiology, treatment effectiveness and safety, and health economic value and impact. Recently, the US Food and Drug Administration and the European Medicines Agency have stated their ambition for greater use of RWE to support applications for new indications, and are now consulting with their stakeholders to formalize standards and expected methods for generating RWE. Pharmaceutical companies are responding to the increasing demands for RWE by developing standards and processes for each stage of the evidence generation pathway. Some conventions are already in place for assuring quality, whereas other processes are specific to the research question and data sources available. As evidence generation increasingly becomes a core role of medical affairs divisions in large pharmaceutical companies, standards of rigour will continue to evolve and improve. Senior pharmaceutical leaders can drive this change by making RWE a core element of their corporate strategy, providing top-level direction on how their respective companies should approach RWE for maximum quality. Here, we describe the current and future areas of RWE application within the pharmaceutical industry, necessary access to data to generate RWE, and the challenges in communicating RWE. Supporting and building on viewpoints from industry and publicly funded research, our perspective is that at each stage of RWE generation, quality will be critical to the impact that RWE has on healthcare decision-makers; not only where RWE is an established and evolving tool, but also in new areas that have the potential to disrupt and to improve drug development pathways.