Skip to main content

Longitudinal monitoring of SARS-CoV-2 RNS on high-touch surfaces in a community setting

Our take —

In this study, published as a preprint and thus not yet peer reviewed, about 8% of swab samples taken from high-touch surfaces in the community (trash cans, grocery stores, liquor stores, crosswalks etc.) in Somerville, Massachusetts were found to be positive for SARS-CoV-2. Given the low concentrations of virus detected and the low transfer efficiency of virus getting from a surface to mucous membrane, risk of infection from touching one of these surfaces was predicted to be low (5/10,000). However, the proportion of positive samples detected on surfaces correlated with subsequent COVID-19 case counts in the community. These data suggest that monitoring the presence of virus on high-touch surfaces may aid in early detection of outbreaks, with a 7-day lead time, though further research is needed with more longitudinal data points.

Study design


Study population and setting

Between April 1 and June 23, 2020, swab samples were collected from high-touch, non-porous surfaces (e.g. trash can handles, post office boxes, crosswalk buttons, etc.) in Somerville, Massachusetts. Samples were collected weekly from 33 unique surfaces at 12 different locations. Sampling locations were observed once per week at the time of sample collection to determine the number of times each surface was touched per hour. The presence of any SARS-CoV-2 RNA was measured using RT-PCR. Presence of virus did not necessarily mean that the virus could cause infection. Authors used a Quantitative Microbial Risk Assessment (QMRA) framework to estimate the risk of infection from contaminated environmental surfaces. QMRA estimates the probability of infection after exposure to a pathogen using simulations that capture realistic ranges of uncertainty for events contributing to infection risk. In this case, the risk of infection was estimated based on the amount of virus on a surface and assuming one hand-to-surface and one subsequent hand-to-face contact. All model parameters were obtained from previously published literature. Pearson correlation coefficients were calculated between percent of positive surface samples and the 7-day case average to examine the association and various lag periods (0-11 days) for case numbers.

Summary of Main Findings

Of the 348 samples collected over nearly 3-months, 29 showed evidence of SARS-CoV-2 by RT-PCR (8.3%), and viral RNA was detected at least once at 10 of 12 locations (never at the post office box or convenience store). Of the 33 unique surfaces, 17 (52%) were positive for SARS-CoV-2 at least once. Three of 29 of the positive samples were detected at sufficient levels to quantify and ranged from 2.5-102 g/cm2 (grocery store door handles on June 16 and liquor store door handle on May 5). Mean touches per hour were highest at the grocery store (n=29) and the gas station (n=20.8). Risk of infection from touching a contaminated surface was low (5/10,000). Weekly percentage of positive samples was most associated with COVID-19 cases 7 days later, and the surface positivity rate explained close to 70% of the variation in cases in the same zip code (r=0.689) and about 55% of the variation of cases in all of Somerville (r=0.548).

Study Strengths

This study measured the prevalence of SARS-CoV-2 on surfaces in a real-world setting, high-touch surfaces at essential businesses, and over time.


These data represent 10 weeks of longitudinal samples and further research is needed to understand how well surface sampling can predict trends in cases. Authors assumed only one hand-to-surface followed by one hand-to-face contact. Although this may be appropriate for surfaces such as door handles, it may be less appropriate for higher traffic surfaces such as gas station nozzles, shopping carts, or items for purchase at stores. Authors discussed the QRMA model being sensitive to changes in parameters, but did not include estimates resulting from these sensitivity analyses.

Value added

This study provided descriptive data on the presence of virus on different high-touch surfaces, further shedding light on the low likelihood of fomite transmission. It also depicted a strategy for how environmental surveillance of surface contamination, that is monitoring the prevalence of virus on surfaces, may be a tool for predicting and understanding community transmission.

This review was posted on: 3 January 2021