Over the years, Clinical Data Management (CDM) has come a long way having its ups and downs, for a while being viewed as a commodity that could easily be outsourced or off-shored. Even today it isn’t one hundred percent clear what CDM is all about. In this brief article I – completely biased – refer to CDM as the function that provides the skills, processes and tools to capture, store, manage, code, check, and present any data related to the health of patients including the meta-data, the derived data and calculated data. The source of the data may be multifold since I do not think CDMs should only be those referred to as working in a CRO, Biotech or Pharma company, mainly in the clinical operations groups. I think one also needs to take a look at other areas where it may be truly beneficial to have CDMs involved.
The classical CDM takes care of creating the case record form (CRF), either on paper or electronically. This already – in an extreme case – can make or breaka trial. If a CRF for a phase 1-3 trial does not adhere to the standards of a specific drug development project, the statisticians and the medical writers will have a very hard time creating the integrated summary of safety, as well as the integrated summary of efficacy. For example, if 9 of 10 trials of a project use the ‘unsolicited reporting of adverse events’ approach, and the 10th study uses a checklist of adverse events (AE), the AE rates will certainly be different and the results very difficult to interpret.
Storing the data in a way so that years or decades later someone, who had no involvement in the management of a trial, still understands what has been done becomes more and more important. That requires the IT department to provide the right storage media, however, the CDM is the one who must ensure the data can still be read and understood. This is an important part of the job taking into account the recent litigations often requiring rather ‘old’ data to be retrieved and made available.
Fifteen years ago, it was absolutely OK for a CDM to be on top of the patient data and not worry about the status of individual fields, outstanding queries, un-coded items, except when getting closer to the status of the locked database. These days, however, many other functions have a vested interest in accessing the clinical trial data on an ongoing basis and have information available on how ‘clean’ the data are. Those functions may be medical monitors, project management, statistical programmers, pharmacovigilance and several other functions. In addition, the desire is to have a locked database available shortly after the last subject in a trial passed his or her last visit. In order to deliver this requirement, CDM has to be on top of every step of a trial on an ongoing basis. This is a challenge and CDM has to identify ways of meeting this expectation.
The data checking becomes more and more important these days, as CDM#s not only track patient data itself (which can already be checked nicely via the EDC applications) but also audit trial data. The CDM can help to identify sites that are usually late with respect to the data entry after a patient visit, do not respond to queries in a reasonable time, change data points frequently, and stick out due to many outliers or patients violating the in- and exclusion criteria. This activity has been neglected frequently in the past, however, has been put on the radar screen more prominently recently. This CDM activity also became important as a result of the introduction of risk-based monitoring, requiring someone with a sound understanding of the data to highlight problematic countries, sites or patients in a trial.
The coding of the adverse events, the medical history and the concomitant medications is an integral part of the broader CDM function. Specialized people understand the rules and regulations in place to ensure text data can be analyzed based on standard thesauri. An unbiased and high quality coding process is the basis of the safety analysis and is an area that CDMs can reach out to if interested in this part of the process.
Finally, and again something that evolved over the past couple of years, CDMs are being asked more frequently to present the data (mainly the process data) of their trials. Someone with project management skills and an interest in performance data can truly shine in this area of CDM, since it is the CDM that is or should be on top of his / her trial.
So, what do I see on the horizon as additional activities for CDMs? In our industry, other functions, like study feasibility management, real life evidence (RLE), market access, genomics, pharmacovigilance and others will need experienced people who know how to deal with data in the broader sense. This may for example mean access to and an understanding of data in electronic health records for feasibility and patient identification. It could also mean the comprehension of huge volumes of data for RLE analyses, retrieval of particular data for market access projects or the huge volume of data in the genomics area. In all cases someone needs to have a sound understanding of the underlying data . If that is not the case, data can easily be mis-interpreted and the conclusions drawn from trash data may be devastating.
The industry also becomes more and more interested in social media data, an area with only little expertise available so far in either the pharma or the CRO industries, at least when trying to use these data for reasonable analyses in the safety or efficacy area. CDMs would be needed to make sense out of the data available and ensure that the completely abnormal data are being highlighted as such.
Finally, from an industry perspective, I think that the mobile health data generated by millions of people eventually will also hit the desk of the CDMs in the industry. One needs to get prepared for it since these data are not necessarily yet as good as one may hope and the CDMs have to start getting involved in order to understand the type and quality of data coming our way. On the other hand, this is a positive challenge and can be a rewarding task.
When looking outside of the clinical trial industry, I believe data managers are being needed in particular in the area of electronic health records in hospitals. This type of data will be used for various purposes in the near future and someone with the understanding of the needs of the industry on one hand and what is doable in a hospital on the other hand would be required desperately.
In summary, despite the valley of tears CDMs went through several years ago, I personally see a bright future for CDMs on the horizon. More data will be generated, however, more data does not necessarily mean more useful data. CDMs can add – using their skills – to making those data usable.