Only 40% of data critical for tracking and controlling coronavirus is being reported in US, finds analysis
The use of accurate, real-time data to inform decision-making is essential for infectious disease control, say experts. The US, however, is failing to report critical information on Covid-19 that could help track and control transmission, and prevent deaths, according to a comprehensive review which concludes that coronavirus data reported across states is “inconsistent, incomplete and inaccessible” in most locations.
The researchers identified 15 essential indicators for effective Covid-19 response — from confirmed and probable cases and testing to contact tracing and hospitalization — but found that overall, only 2% of the exact essential indicators were reported, 38% of indicators were reported in some way, but had data limitations or did not stratify data adequately, and 60% of indicators were not reported in any way. The majority of data missing is related to testing and contact tracing.
"In our review of public data dashboards from all 50 US states and the District of Columbia, we found that only 40% of essential data points are being monitored and reported publicly. More than half the essential information — strategic intelligence that leaders need to turn the tide against Covid-19 — is not reported at all. There is high variability in the data presented, the way it is presented, and the level of information shared," says the assessment by Resolve to Save Lives, an initiative of the global public health organization Vital Strategies. It is led by Dr Tom Frieden, the former director of the US Centers for Disease Control and Prevention (CDC).
In the absence of a clear national vision, strategy, leadership or organization, it is crucial to establish standardized, timely, accurate, interlinked, comparable and informative dashboards for every state and county in the US, write authors, who collected the information from June 30 to July 14.
"Despite good work done in many states on the challenging task of collecting, analyzing, and presenting crucial information, because of the failure of national leadership, the US is flying blind in our effort to curb the spread of Covid-19. By revamping our approach and using data to empower people to know their own risk and ensure accountability for decreasing that risk by improving our response, we can get our kids back to school, ourselves back to work, accelerate economic recovery, and save tens of thousands of lives," says Dr Frieden.
Over five months into the Covid-19 pandemic, not a single state reported the turnaround time for diagnostic test results, a crucial indicator of program effectiveness. "If test results are delayed beyond one or two days, they become much less useful, as those who are infected will have passed the period of peak infectivity and will neither be able to be rapidly isolated (unless all who are tested are isolated until result return) nor warn exposed contacts so that they can quarantine before they become infectious. All states should report diagnostic test median turnaround time (specimen collection to test report), by week," says the report.
Data on Covid-19 testing is critical for putting observed case trends in context and understanding the risk of transmission. For example, there may be a low number of cases detected in an area, but if testing rates are also low, transmission risk may still be high, as many cases remain undetected. More than 90% of states report information on trends in the daily number of PCR tests (for active infection) performed. But among the states that make information on testing public, reporting is inconsistent.
Some dashboards report numbers of tests without specifying whether this refers to PCR testing, antigen testing (for screening), antibody testing (for past exposure), or a combination of these tests. "For many states, it is unclear if the number reported represents tests conducted or the number of people tested. These details are critical to interpreting testing data correctly and can lead to erroneous conclusions or inaccuracies when national or regional estimates are produced," it says.
Rapid isolation of cases and quarantine of contacts is the bedrock of outbreak control, especially in the absence of effective medication and vaccines. The review, however, says that currently, state reporting on contact tracing is ‘abysmal.’ Of 156 critical pieces of information on contact tracing programs, only three (2%) are reported in some way and zero in the optimal way needed. Only two states report data on how quickly contact tracers were able to interview people who test positive to learn about potential contacts.
"Just eight states reported data on the source of exposure for people who test positive; cases coming from unknown exposure signal much higher risk from undetected community transmission. And not a single state reports on the percentage of cases arising among people that are under quarantine after being notified of their potential exposure by a contact tracer—the single most important indicator of the effectiveness of contact tracing. Some other states indicate contact tracing metrics are coming soon, but provide no estimate of when that might be. Without this information, states and people who live in them have no way to assess the effectiveness of the response in their state," says the team.
The report recommends that all states should report key contact tracing performance indicators, including the percentage of new cases epidemiologically linked to at least one other case by date, stratified by whether part of a known outbreak or not; time from specimen collection to isolation of cases, by week; percentage of cases interviewed for contact elicitation within 48 hours of case specimen collection, including all people with positive tests who reside in the jurisdiction, by week; and percentage of new cases from among quarantined contacts, by week.
All states have a Covid-19 data dashboard, but because state dashboards were developed independently, no two are identical in terms of information presented, usability or look. While cases are a basic indicator of any disease program, the review says case reporting is "surprisingly inconsistent" across state dashboards. Timely information is essential to inform public health action, but 20% of state dashboards did not report same day data by 5 pm local time, and two states displayed data from two or more days prior. The researchers recommend that all states should report data the same day it is collected. While reporting new cases by date — which indicates disease progression over time is an epidemic curve — 8% of states do not display this figure directly on their dashboard.
Nearly 40% of states do not present any information other than new or cumulative confirmed cases. Reported case counts include several types of cases (probable, or not yet lab-confirmed, confirmed, recovered), but some states do not specify case type, so it is unclear what is being reported. "There are also variations in how reported cases are assigned a date (from date of specimen collection (preferred) to date of illness onset, to date reported), but some state dashboards do not make clear which type of date is used. All states should report new confirmed and probable cases and per-capita rates by date with a seven-day moving average. Moving averages of less than 7 days do not capture weekends at all times, and weekends tend to have systematically lower case counts," the authors explain.
(Getty Images)
Currently, only 18% of states are reporting on flu-like illnesses on their Covid-19 dashboard (despite widespread collection of this information) and only 37% of states are reporting Covid-like illnesses — both important early signals of potential coronavirus spread. One third of states do not report any data on outbreaks in congregate facilities such as nursing homes, homeless shelters, correctional facilities, leaving communities without the information necessary to protect these vulnerable populations.
The review calls on all states should to most essential indicators such as cases, testing, hospitalizations, and deaths by key demographic indicators including age, sex, and race, and ethnicity. This data should be reported weekly, not just cumulatively so that important recent patterns can be easily identified, they add.
"States should follow basic best practices for data sharing including reporting data the same day it is collected to inform timely risk assessment and action, linking to county/city data dashboards so that the public can easily see what is happening in their county and neighboring counties, and Linking to county/city data dashboards so that the public can easily see what is happening in their county and neighboring counties," says the report.
The authors say that this could be extremely challenging, requiring changes in the current data system, new data agreements, and engagement with entities outside the public sector. "It will also require additional financial resources and staff, but the cost of not investing in better information is far exceeded by the cost of inaction and the prolonged health, economic and social impacts of Covid-19," the report warns.