The Future of Public Health Surveillance
Dr. David V. McQueen talks about the future of public health data collection and surveillance and how epidemiologists must position themselves to maintain their future relevance.
Could you describe your achievements in the field of health surveillance?
My main work in health surveillance began when I was a professor at the Johns Hopkins School of Public Health back in the late 70s and early 80s as a consultant for the US Centers for Disease Control and Prevention (CDC). At that time the CDC established a team of consultants to develop the strategy for what would become the Behavioral Risk Factor Surveillance System (BRFSS). The idea was to develop a system where each state would collect data related to risk factors for chronic diseases. In the beginning it was quite simple. We took the leading causes of death from chronic diseases such as cancer, heart diseases, diabetes, etc. and looked at what were the individual risk factors related to these diseases, such as smoking, diet, physical activities, alcohol use and so on, and asked questions related to that. We then trained employees at the state level to collect data through telephone interviews, which was quite innovative compared to visiting houses and/or mailing out questionnaires.
Shortly thereafter, I became professor and head of a new research unit in health and behaviour change (RUHBC) at the University of Edinburgh in Scotland. The Scottish Government wanted to find out why Scotland had such dreadful chronic disease outcomes. The experience from CDC and the development of BRFSS led to a move from behaviour risk factor surveillance, to surveillance of lifestyle and health and to the introduction of computer assisted telephone interviews (CATI). Since personal computers were now available, we could hire interviewers to collect data on the phone and input it directly into a PC. This technological breakthrough enabled us to collect data “continuously”.
Historically, one of the problems with data was the variation in personal behaviour related to time in the week, time in the year and so forth. By collecting data continuously we would minimize this data error. But we went further by collecting data systematically, analyzing and reporting the data systematically. This was a methodological breakthrough from surveys with infrequent data collection patterns. One of the main problems with surveillance data — particularly in risk factor surveillance — was that it wasn’t current. One collected the data and it could go two to three years before the data was published. We wanted to solve this problem.
After 10 years in Scotland, I came back to CDC and I was asked to be director of the BRFSS. I led that for several years and thought about spreading it to other countries. We started in 1999 with the first meeting that became a bi-annual meeting bringing together people from different countries and assisting them to develop similar programs. These particular meetings ultimately became the World Alliance for Risk Factor Surveillance (WARFS) and we decided to join with the International Union for Health Promotion and Education (IUHPE) as it was important to link risk factor surveillance to health promotion and to the use of the data for policy and ongoing programs.
You’ve retired from the CDC - are you still active in the global community of surveillance?
What motivates me is that the vision I had over the years — building useful and continuous data collection and connecting with people that are interested in this idea — that vision is not complete. There is a lack of people who are collecting data that tries to look at the causes of the causes in chronic diseases, but I think there is also a strong need for much more and better analysis. I think that most of the time we just write reports with cross tabulations by single variables – e.g. how many people smoke rather than trying to build analytic systems that look comprehensively at how these variables fit together. I think that that lack of in-depth analysis, and tying the usefulness of the data for policy and health promotion as the end goal for surveillance, is still work in progress.
Where is public health heading in general?
I think public health at present is in retrograde motion. Public health was moving along quite well, but due to the events happening in the world public health is retreating back to its hygiene base with a focus on individual disease causality. I think this has been accelerated by the ebola epidemic, bio terrorism, influenza outbreaks, etc. Those events get attention and avert people (and resources) away from other areas. So I think that the chronic disease risk factor surveillance area is still in the same struggle for recognition.
A few years back, there was so much interest in the social determinants of health after the Marmot report came out — and for a moment everybody was paying attention. There are probably many factors for the loss of this momentum. Infectious disease outbreaks do get much attention, but I also think that the government’s austerity programs have got public health to retreat to its comfort zone — infectious diseases. If you are going to cut programs, you will cut programs that are at the edge of the public health sector. This is why health promotion currently suffers so much. It appears to people as a luxury, but I would argue that it is the opposite.
What are the biggest challenges epidemiologists face today? How is the profession preparing for these new challenges?
I think the profession of epidemiology is going through a lot of changes. The historical success of epidemiology is in the description of the prevalence of disease — that has always been its big focus. The biggest challenge is to move away from that simple description, as it doesn’t tie into health promotion and the broader social context that is necessary to make epidemiology more meaningful. There have been a lot of good efforts over the last 5-10 years to develop social epidemiology — integrating epidemiology into social concepts. But they are very limited and still not mainstream as far as I can see. For me the main challenge is that epidemiology has become “over-biomedicalized” and needs to be more about public health and a bit less about clinical medicine. It doesn’t have a good understanding of health and disease in a broader sense. There is no good epidemiology epistemology about what the field is and therefore it retreats to more simplistic approaches.
Are we in public health focusing on the right things and using the right methodologies?
To a limited extent I think we have done quite well. There is, however, a need for a radical overhaul of what we ask about, and how data are used and thought about. The sort of problems that come up — even if you are looking at the social determinant literature — talk about issues like poverty and inequity and so on. But the analysis of data that comes from this is looking at risk factors and retreating back to looking at smoking, drinking, sedentary behaviour, etc. This is where the science is — the comfort area. If you think that the ‘causes of the causes', are the really important things, then those are the type of data you need to look at and concentrate on.
Do you think that the data you define is readily available and you really have to work hard to find the other data?
Yes, I think so. But again, so much is involved with training. If you think how people are trained in public health — they have traditionally been trained in biomedicine and from that perspective it is not surprising that this is a comfort zone when dealing with diseases and direct outcomes — even behaviour as an outcome. We can think about the causes of behaviour — what behaviour leads to causes of disease outcome — but we don’t. We tend to not go there. We really do need to rethink how we collect new data.
What impacts will emerging technologies have on public health risk factor surveillance?
I think they will have enormous impact. There is no doubt about that. If you look back to my personal history and the development of computers and the patchy approaches to collect data and to analyze them, that is an example of technological impact that really made a difference in the way we made surveillance of risk factors. I think what is happening now with all the new devices that are available and the use of the internet and so forth, is really an area which will have terrific impact on how we collect data, how we look at data, what constitute data, and how to rethink what it is that we are looking at. So yes, I can just say if I were not retired and if I were working in this field full time this would be the area I would really concentrate on — because this is the future.
Who is working in this area?
Some of the people who are working with WARFS are starting to discuss this topic. I know that people at the CDC are worrying about this, and people working in behaviour risk factor surveillance are thinking about the use of such things as the internet. But it really could mean a totally different way of understanding “what is a sample?” This is one of the problems and requires a huge rethink. If you take an area like sampling, this development has enormous implications because sampling used historically is based on where people live, if they are in a household and all those kinds of things. And now suddenly sampling is coming out of people’s wearable devices and the social media they are using. What constitutes a “sample” seems like a fundamental question. I don't know the answer to this question, but I am hoping that there are people that are really considering this in detail. And I am hoping that in our meetings in WARFS and other organization we are really addressing this issue.
Where do you look for information on new developments?
I think it follows from what we have discussed before. I think that people who are working in this area need to get out of the public health biomedical box and look at efforts in the social sciences — in the so-called softer sciences — and looking at other potentially relevant approaches. We have to move away from seeing that the solution is going to be found in the traditional public health boxes. That requires different kinds of people to get involved in surveillance.
What advice have you got for CARRFS?
I think that organizations such as CARRFS, WARFS, and the Americas’ Network for Chronic Disease Surveillance (AMNET) are very good cutting-edge organizations. The problem is that they are on the periphery and not recognized as much as they should be by the established public health community. The ideas coming out of these organizations need to get into the mainstream of the public health sector. These organizations need to be recognized by the established public health community such as the national agencies, the CDCs, the European CDCs and the World Health Organization (WHO). Many of these organizations tend to think that they have the expertise. I would, however, argue that it is the institutions like CARRFS, AMNET, WARFS, etc., the organizations that are actually doing surveillance, that have the expertise and need to be listened to. The importance of these organizations has to be recognized by these other organizations themselves as well as the global and national public health institutions. <>
By Jostein Algroy