Dear Susan.

I agree with you that there is a little interest to “use” epidemiologist in research conducted by other clinical specialists also in the public health area. I think this is due to several reasons. The first one is, that the standard curriculum in epidemiology has little information about non-classical statistical methods of data analysis (such as clustering, neural networks, and so on) - as far as I can say – in Poland. Moreover, modern software in which one may just click to get results of the analysis (such Stata, Statistica or Systat) give a false impression that epidemiologists (biostatisticians) are no more useful. This is of course a huge mistake (hardly ever such persons check necessary requirements or calculate the required size of the study group in advance). The last but not least problem is, that data scientists who worked with clinicians need to have quite extensive knowledge of biological and medical background in developed areas, which may sometimes be a challenge for those without adequate curriculum and passion. All of that leads to this situation, that there exists a misunderstood idea of cooperation between epidemiologists (data scientists) and clinicians.

There is an increasing demand for clinical epidemiologists. If both the data scientists and clinical epidemiologists work together, there will be dramatic improvements in health care in many developing countries. lalitha kabilan

Dear Susan.

I agree with you that there is a little interest to “use” epidemiologist in research conducted by other clinical specialists also in the public health area. I think this is due to several reasons. The first one is, that the standard curriculum in epidemiology has little information about non-classical statistical methods of data analysis (such as clustering, neural networks, and so on) - as far as I can say – in Poland. Moreover, modern software in which one may just click to get results of the analysis (such Stata, Statistica or Systat) give a false impression that epidemiologists (biostatisticians) are no more useful. This is of course a huge mistake (hardly ever such persons check necessary requirements or calculate the required size of the study group in advance). The last but not least problem is, that data scientists who worked with clinicians need to have quite extensive knowledge of biological and medical background in developed areas, which may sometimes be a challenge for those without adequate curriculum and passion. All of that leads to this situation, that there exists a misunderstood idea of cooperation between epidemiologists (data scientists) and clinicians.

It does seem like something that is entirely possible, and I wonder where "Epidemiologists" such as me would fit in then!

I don't see "increasing IT based Business Intelligence being deployed in healthcare" as a threat to epidemiologists, as much as a threat to "epidemiologists who don't know how to carve their space in an ever-evolving technological World". If epidemiology has evolved to the point that we need more and more epidemiological-intelligence, then it is the role of the epidemiologists to speak that language, and be themselves the ones who bring in the data scientists. The data scientists can't go far without the help of epidemiological thinking, the same way an epidemiologist nowadays can't go far without at least understanding hoe these technologies can help their work. An epidemiologist needs not become a data scientist him/herself, but one needs to at least understand what language they speak and be able to communicate/collaborate with them, in order to profit form their invasion of healthcare, rather than lose space with it.

There are already epidemiologists and biostatisticians who specialize in the analysis of very large datasets, particularly the Medicare databases. Modelers of disease spread among populations are using anonymized cell phone databases. I think the answer will be Yes, the two disciplines studying human behavior will certainly merge methods and analytics.

"Epidemiologist" Old wine in a new bottle "Data Scientist", as evolution takes it...

Snow collected his data himself, going from door to door to find out whether the households where cholera struck had collected their drinking water from the Broad Street pump. Data collection has changed greatly today, now we are talking about big data (data in the range of Yotta bytes) which is being collected routinely through robust technologies in the healthcare arena. Epidemiologists/bio-statisticians are best equipped to handle this, as they have been doing it for ages, since John Snow.

Thankyou for all your thoughtful comments. One of the articles that prompted me to wonder about this was a recent editorial in J Clin Epi, “Has the time arrived for clinical epidemiologists to routinely use 'routinely collected data'?”

(https://www.researchgate.net/publication/236938226_Has_the_time_arrived_for_clinical_epidemiologists_to_routinely_use_%27routinely_collected_data%27), which suggested to me a disconnect between the Data Sciences already developing technologies to utilise routine data sources and the classic epidemiological sciences, which are perhaps more reticent to use ‘dirty data’. Developments such as the RECORD guidelines for use of medical records data will hopefully help close the gap and provide the methodological rigor and caveats around the use of this data to really utilise this growing resource and harvest the insights that it must surely hold.

Article Has the time arrived for clinical epidemiologists to routine...

Hello, I hope I have understood your question and I'd like to share a data:

Haynes, an Epidemiologist , is the author of the most used methodological filter for data search in Pubmed. The rol of epidemiologists, biostatisticians, and librarians is today (in each case) very changing. But above all I believe there is a very big opportunity to work in an integrated way with IT professionals. Don't you think is just about time to become more integrated in common problems and solutions? This would probably get more skills and professional expertise.

See what I say: http://www.ncbi.nlm.nih.gov/pubmed/clinical

Greetings

Thanks again, there seems to be a lot of agreement that these scientific fields need to move closer together to produce greater effect. I see this occurring at a higher organisational levels such as government and eg Centres for Medicare & Medicaid, etc. and large premium Healthcare Organisations (eg Mayo, VHA). However the diffusion of this type of work to the smaller organisations and clinical service frontlines seems to be SLOW. So I am going to post a new question to ask for your advice and thoughts: What are the key issues that need to be addressed by a program to improve use of clinical data for health service management and clinical quality improvement? I theorise that it requires both the implementation of technological tools to provide better and wider access to the disparate stores of routinely gathered clinical data (eg EMRs, specialised clinical information systems) AND concurrent training of the broader clinical workforce in applied epidemiology/stats/informatics/data sciences to know how to start better using these resources. What do you think? (new question posted at https://www.researchgate.net/post/What_key_issues_need_to_be_addressed_by_a_program_to_improve_use_of_clinical_data_for_decisionmaking_in_health_service_management_and_clinical_QI)

Susan, you certainly ask interesting, important questions. In my years working with the US CDC and now working with the Taskforce for Global Health (TEPHINET) and Skoll Foundation Global Threats Fund, I am seeing this future arriving. We are training epidemiologists to be data scientists. Formal surveillance data used to be a one-way street from the field to the central office. "Edge" actors in the system had little access to decision-making data. That is changing. Projects such as Google Flu Trends and other social-network oriented informal surveillance data efforts are placing the data and control in the hands of people on the ground in ways not possible a few years ago. I am seeing this prompting an opening of the formal surveillance data too as well as updating to the national and international standards for field epidemiology training curricula. Basically, I do not see data scientists replacing epidemiologists. Rather, I see epidemiologists taking on the knowledge base and tools of data scientists. This is true even for in-house clinical process and quality control managers.

http://healthmap.org/ddd/