The latest data from our athenaResearch team shows a nationwide decline in the flu for a fourth consecutive week and a 25% decline from this season’s peak rate, which occurred the week ending December 28, 2013 (see Figure 1). If this trend continues, this week’s update may serve as our last flu report for the 2013-2014 season.
While the national rate declined, there are some states where our data showed an increase (or no decline) in flu, including the New
England region, particularly Connecticut, Massachusetts and Vermont. Our data shows modest increases each week in those states, but with flu levels far below last season’s peak (week 4 rates of 0.6%, 0.4% and 0.7% rates respectively for the three states, compared to national rate of 0.7% for week 4 and 1.1% three weeks earlier).
In general, the flu is running at levels roughly 30% lower than last year’s national level, with some states seeing extremely benign flu seasons relative to last year. Centers for Disease Control spokesman Erin Burns said last week, “I don’t think we can say yet that the flu season has peaked,” but our team here at athenahealth is encouraged by the sustained decline.
The athenaResearch Perspective
Now that my team is at the tail end of our first disease surveillance – the 2013-2014 influenza season – I wanted to share some of our experiences using a cloud-based IT infrastructure as we have to track health-related data and report findings.
Using the Cloud to Monitor Disease
As described in previous posts, athenahealth operates from a single instance, cloud-based infrastructure. This means that all of our clients access the same software platform and use common clinical, operational, and business definitions. All documented care is immediately added to a secure, hosted database that our research team can access for analytic purposes. This gives us highly granular, real-time visibility into clinical practice patterns and patient treatment for our more-than 47,000 providers across the country. These are the data assets our team has applied to monitoring flu diagnoses over the past 13 months, and that we’ve looked to for other recent studies, such as behavioral health in pediatrician visits.
athenaResearch and the CDC
Our approach to disease surveillance differs from that of the CDC. The CDC measures flu based on weekly reports from roughly 1,800 outpatient providers across the country who report on the frequency of influenza type illness (ILI) in their population. They also report on flu lab results, strains and deaths.
In contrast, athenaResearch reports on the percentage of patient visits that included a diagnosis of flu, documented as an ICD-9 code on a claim. Each week, we collect data from visits to over 16,000 primary care physicians (PCP), nationwide. Although our methods differ from those at the CDC, our trends have generally mirrored theirs with high consistency.
Cloud-based Surveillance Provides a Safety Net for Public Disease Surveillance
During last fall’s 16-day federal government shutdown, the CDC was forced to furlough 68% of its work force, cutting their flu monitoring efforts. Because a flu outbreak in October is not unheard of, we saw an opportunity to help during this period of exposure, and published flu trend reports for the duration of the furlough. Fortunately, flu incidence during this period was very low; but, had a flu outbreak occurred during the government shutdown, we would have detected it and reported on it, both here on our athenahealth blog, directly to athenahealth providers, and throughout the Epocrates network via alerts. The larger lesson: Private sector, cloud-based disease surveillance activities provides a degree of valuable redundancy to public sector approaches. A safety net, of sorts.
In addition, cloud-based EHRs allow for situational awareness, where physicians can be alerted to pathogens circulating in their communities and can educate patients accordingly. Part of our roadmap at athenahealth calls for us to alert physicians when flu or other communicable diseases start presenting in their communities. Imagine a PCP alerting an at-risk patient who has not received a vaccination that the flu is projected to hit their area in three weeks based on neighboring states – still enough time to receive a vaccine and be protected.
A More Timely Source of Data
At athenahealth, our technology platform enables us to report on claims submissions or clinical documentation in real time, and generate daily data with a one-day lag, when appropriate. That is the power of cloud technology. Comparatively, many public health reports provide weekly data with a seven-day lag — and depending on the day of the week, public health reports may provide data from7-13 days prior. (For diseases with large public health implications, that time frame can be compressed, but data may be incomplete.) Cloud-based disease surveillance can “fill in the blanks” between regular weekly reports and reportable event systems, and can also provide data on diseases that aren’t included in regular reporting.
A More Robust Database for Projections
A cloud-based database contains a multitude of patient characteristics — address, location of diagnosis, demographics, clinical characteristics, lab values — that may be particularly useful in tracking the flu. When combined with other data streams (like flu look-up data online or flu mentions on Twitter), projections of flu incidence are likely to be more robust. We are currently speaking with several research groups about providing them data feeds that will help them refine and segment available flu incidence projections.
Supporting Local Public Health Agencies
In cases where athenahealth serves more PCPs in a particular region than an existing disease monitoring network, we may be able to provide useful supplemental information to local public health authorities.
As an example, we serve a large number of physicians in the Columbus, Ohio region, and supply the Columbus Department of Public Health with weekly data that goes beyond what they’re receiving through other sources. Although these activities have been just for the flu, we envision other valuable opportunities for cloud-based IT companies to supplement the information available to public health authorities. Monitoring Lyme disease, pertussis, respiratory syncytial virus, chickenpox, and GI infections could all be valuable.
If you work with a public health organization and would be interested in a proprietary data feed from athenahealth, we would love to speak with you. Please email our Director of athenaResearch, Iyue Sung, for details: email@example.com
In a mature cloud-based environment, identifying disease surges or declines is just the first step. Well-curated data sets and reporting tools can provide a platform for answering a series of questions: Once flu is diagnosed, how do physicians behave? How often do they prescribe antivirals or antibiotics, and to what type of patients? For patients with neurological disabilities, what are the differences in flu rates for vaccinated and unvaccinated patients? We are working to build a system that effectively identifies disease surges, and then drills down to distill useful insights.
January 27, 2014