spacer
home > ict > winter 2013 > under surveillance
PUBLICATIONS
International Clinical Trials

Under Surveillance

On-site monitoring is a resource-intensive component of clinical trials, contributing to over 30 per cent of their costs. Moreover, it is estimated that only about 50 per cent of the monitor’s time is actually spent on monitoring activities, while the rest is spent on travel, logistics and reporting.

As a result – and in view of the low R&D productivity and major cost pressures on the pharma industry over the past decade – it is hardly surprising that there has been an increased focus on reducing the cost of clinical trial monitoring by making the process more efficient.

Centralised, remote, risk-based and adaptive monitoring are effective strategies which reduce the need for on-site monitoring visits and source data verifi cation (SDV). Indeed, both the US Food and Drug Administration (FDA) and the European Medicines Agency are urging greater reliance on centralised monitoring practices to identify when on-site monitoring is truly required, and how it can be optimised using metrics collected by centralised monitoring methods.

The FDA’s Guidance for Industry Oversight of Clinical Investigations – A Risk-Based Approach to Monitoring, released in August 2013, recommends a quality risk management approach to clinical trials. It encourages greater use of centralised monitoring practices and other alternate monitoring approaches, which may be suited as sponsors develop risk-based monitoring strategies.

Reliable Data and Patient Safety

The overarching objective of clinical trial monitoring is to improve the availability of reliable data from trials which are used to assess and endorse safety and efficacy of new treatments. Monitoring should enhance the quality for study participants and also for future patients, who will receive the medicinal products approved on the basis of regulatory submission of data from clinical studies.

The default, resource-intensive monitoring, which traditionally consists of on-site visits and 100 per cent SDV, is no longer considered an optimal method. This is because it cannot achieve the primary objectives of monitoring without other complementary approaches working in tandem. With this in mind, most companies now adopt a risk-based approach to determine the optimal monitoring strategy for each study.

Risk-Based Approach

Protocol complexity, study duration, investigator experience, recruitment rates and the extent of electronic data capture are a few key considerations used to assess the risk involved in the study and to determine how a centralised monitoring approach can be used.

A major determinant of protocol complexity is study design. Studies can be assigned risk ratings a priori, depending on whether only non-invasive procedures are used (lowest risk – for example, non-interventional/ observational); only approved treatments are used (mild risk – for example, Phase 4); whether it is a Phase 3 trial with a new treatment and/or a new indication (moderate risk); or whether it is a Phase 1 or Phase 2 trial of a new treatment (high risk).

An optimised monitoring strategy can be determined for every new study on the basis of these risk ratings. Modification and customisation of the strategy may be required for individual investigator sites based on key risk indicators (KRIs) related to aspects of study conduct, patient safety, treatment compliance and data management. It is important to identify the KRIs upfront so that the metrics required to evaluate and monitor them can be generated and reported on a regular basis.

A hybrid monitoring approach is determined by factoring in the risk ratings and KRIs. The monitoring plans may sometimes include extensive on-site monitoring, but mostly include reduced monitoring (risk-adapted, or on a random sample of centres, patients or outcomes), or targeted monitoring (based on KRIs and statistical monitoring). Using this advanced approach, the quality of study conduct is monitored through:
  • Training, mentoring and oversight of site staff
  • Data entry checks and discrepancy management
  • Central and on-site monitoring of data
  • Planned checks carried out by data monitoring committees (DMCs)
CSS Adds Value

Central statistical surveillance (CSS) is a critical element of central monitoring. This technique uses statistical tools to identify errors, outliers and abnormal trends/ patterns in clinical trial data, and thus provides effective triggers and leads for targeted monitoring visits.

Sources of errors are study design, study procedures, case report form (CRF) design, data recording, data analysis and inference. Errors can be random or systematic. Random errors may impact the statistical power of the study but may not result in bias, while systematic errors will most certainly result in biased conclusions.

Errors could be unintentional (for example, an unknown issue with calibration of an instrument); due to lack of attention to detail (data not transcribed correctly from the source to the CRF); or due to a lack of understanding (unclear about how dose titration details are to be captured on the CRF). In rare instances, there could be deliberate errors committed with the intention to fabricate or falsify data.

Statistical Checks

The main principles underlying the use of statistical methods to detect errors, outliers or trends are:
  • Clinical trial data generated by application of a standard protocol and the same CRF at all participating centres is highly structured in the way it is collected, grouped and analysed
  • The multivariate structure and/or time dependence of the variables is sensitive to deviations and easily detected by statistical tests
CSS is able to detect data issues that go undetected in SDV and on-site checks, and considers every piece of information entered in the CRF as potentially indicative of quality, rather than being restricted by predefi ned KRIs. Statistical checks are performed to check randomness (fi rst digit preference, rounding), plausibility (correlation structure, outliers, dates in range) and comparability (between treatment arms, between centres or any other covariates of interest).

Procedures and Tests

Basic statistical procedures and tests such as Chi-square test, t-test and F-test are used to compare the distribution of all variables of interest across centres, for the purpose of identifying any outlying centres or observations. Study design and CRF design issues could also be detected from such analysis. For example, low variability in data across visits (indicated through the F-test) may lead to suspicion that the same observation is being entered without actual measurements being taken at each visit. This may suggest that the CRF has too many fields, and its size may need to be reduced.

Plausibility checks (for correlations, outliers) may require appropriate plots and graphs, followed by statistical inferential procedures. Model-based approaches are required to check comparability across centres. Tests on proportion of outliers, means, within patient variances and so on are used to generate a high-dimensional matrix of p-values, and statistical methods are used to identify outliers.

Data fabrication or falsification done with an intention to hide missing data, or to make the results look more favourable for a particular treatment, may be detectable through treatmentby- centre comparisons from appropriate statistical models that are fitted to the data.

However, CSS relies on sufficient data being available to be able to detect abnormal trends. As a result, it may not be effective in trials that have several centres with small amounts of data, or at the beginning of a trial before suffi cient data are accumulated.

Powerful Tool

In summary, an efficient and effective monitoring strategy is a combination of targeted site visits, primarily for training, mentoring and supporting site staff; remote assessment through incident alerts, tracking systems and statistical analysis; and trial oversight through steering committees and DMCs.

CSS is a powerful tool which can identify unexpected or strange data patterns, making on-site monitoring much more targeted and effective. Its growing importance is indicative of the crucial role of statistics and programming in effective clinical trial monitoring and hence the need for the sponsor to ensure availability of statistical expertise while planning resources for monitoring.

If the sponsor is managing the trial in-house, adequate statistics and programming resources will have to be planned for each study. If the trial is outsourced to a contract research organisation, the sponsor will have to ascertain that organistion has the capability to perform CSS, or they may have to use another vendor for this activity.


Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:
0
     

There are no comments in regards to this article.

spacer
Chitra Lele is Chief Scientific Officer at Sciformix, with over 20 years of experience in the healthcare industry. She has been part of the company’s leadership from its inception, and has been instrumental in establishing and growing the organisation. Prior to Sciformix, Chitra was Executive Director responsible for Indian operations of Pfizer Global R&D. With a PhD in Statistics from Stanford University, her prior experience includes work as a biostatistician in cancer epidemiology at both Stanford and University of California.

Samyuktha Ajay is Director of Clinical Development at Sciformix. She is a pharmacy post-graduate from Mumbai, India, with more than 20 years of experience in various clinical research roles in multinational and Indian pharmaceuticals companies and contract research organisations.
spacer
Chitra Lele
spacer
spacer
spacer
Samyuktha Ajay
spacer
spacer
Print this page
Send to a friend
Privacy statement
News and Press Releases

MedPharm announce expansion of partnership with Palvella Therapeutics in developing a new treatment for the debilitating rare disease, pachyonychia congenita

MedPharm Ltd have announced the expansion of their partnership with Palvella Therapeutics, Inc., a Philadelphia-based biopharmaceutical company focused on developing and commercialising therapies for debilitating, rare genetic diseases. To date, MedPharm has employed its world-renowned specialist formulation expertise to support Palvella’s development of a novel, high-strength rapamycin topical formulation for application to the skin (PTX 022) as a disease-modifying treatment for pachyonychia congenita (PC). Most recently, MedPharm has made arrangements to manufacture the clinical (IMP) batches for use in Palvella’s upcoming Phase 2/3 clinical study.
More info >>

White Papers

PROTEIN:PROTEIN INTERACTION ANALYSIS

IBA GmbH

Groth et al., 2007 and Jasencakova et al., 2010 Protein:protein-interactions (PPI) govern almost all important processes in living organisms. Thus, their rapid and accurate determination and investigation is a major challenge in life sciences. The extremely efficient and fast One-STrEP (Twin-Strep) system is suitable for the isolation of functional protein complexes and subsequent mass spectrometry analysis leads to the identification of protein complex components (Junttila et al., 2005) as it is shown in the following studies.
More info >>

 
Industry Events

Clinical Innovation Partnerships

11-12 September 2019, Hotel Palace, Berlin

Clinical Innovation Partnerships is the only event for senior clinical operations and procurement professionals driving innovation in their clinical trials. It is where the pharma, CRO and vendor clinical trials ecosystem meet to drive innovation in digitalisation, tech and partnering to improve patient experience and reduce trial cost and time.
More info >>

 

 

©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement