How sewage surveillance lit Hong Kong’s pandemic path

Picture courtesy EPD, Government of HKSAR

Dr Samuel HK Chui explains how using wastewater surveillance in Hong Kong to help manage the spread of COVID-19 was like driving with the lights on.

Despite being a busy international aviation hub with a very crowded living environment, Hong Kong managed to maintain a very low infection rate in the first four waves of the COVID-19 outbreak. This was down to the strong public awareness of personal hygiene and effective implementation of city-scale non-pharmaceutical intervention measures. However, the unexpectedly high transmission rate during the onset of the fifth wave of COVID-19 from early 2022 put the city at risk.

On 1 February 2022 there were about 100 new cases per day. By comparison, on 3 March 2022 this increased to 77,000 new cases per day at its peak – 1% of the city’s total population.

Sewage sampling in action

During this time the city exhausted its local clinical COVID-19 diagnostic capacity, resulting in public anxiety due to uncertainty regarding the threat of undiagnosed people spreading the disease in the community. To help find ways of controlling the chain of transmission quickly and in a cost-effective manner to keep social disruption to a minimum, the Hong Kong Special Administrative Region (HKSAR) government implemented sewage surveillance to identify high risk areas at city-block-level via analysis of spatial and temporal variations in viral data.

Work-flow of various teams in the city-wide surveillance programme

Resources were prioritised to introduce localised risk-based compulsory clinical testing using polymerase chain reaction (PCR) tests.

This non-intrusive surveillance strategy helped Hong Kong tide over the sudden arrival of this peak wave of COVID-19, with new cases progressively dropping to less than 1,000 per day in mid-April 2022 and to 237 cases per day on 22 May 2022.

Prediction of the number of new cases

Sewage surveillance programme

Hong Kong’s sewage surveillance programme was the first city-wide, real-time interactive sewage network-based virus control programme reported in open literature. This evolved from a ‘practical prototype’ that was scientifically proven during the fourth wave of COVID-19 to uncover hidden cases to provide early warning signals of outbreaks and spread in the community; monitor infection clusters for intervention actions; and track the development of the epidemic. The programme was a collaborative effort involving more than 400 individuals with diverse specialities, and was achieved through synergy throughout government, academic and private sectors to scale up the prototype into a full programme in just a few months.

Viral concentration and the number of confirmed cases

How it works

While surveillance programmes worldwide typically collected sewage samples at sewage treatment works, Hong Kong’s sewage surveillance programme is sewage network based. This approach could better capture the spatial distribution of COVID-19 virus signals across the territory, and at the same time avoid random errors arising from the unique centralised sewage treatment network in Hong Kong. Leveraging Hong Kong’s full and well-documented sewerage reach of 93% of its population, the multi-disciplinary team established a comprehensive surveillance network with 154 stationary sampling sites, covering six million people (over 80% of the city’s population), with each sampling site covering an average population of 40,000 people, carrying out intensive sewage monitoring for COVID-19 in a cycle of once every two days.

GIS showing locations of sewer sheds in HK

These stationary sites are manholes strategically selected based on population coverage, feasibility of sampling and upstream signal tracing, impact on traffic, and clearance of hospital effluent and quarantine facilities. A list of suitable upstream manholes (more than 2,000) was also prepared for additional sewage collection through upstream tracing to narrow down the area of concern from a large sewershed to a city block or even a single building, enabling small-area compulsory testing orders to reduce disturbances to the community and enhance efficiency.

Laboratory testing capacity was also one of the key success factors in coping with the fifth wave of COVID-19. To secure sufficient testing capacity to analyse more than 100 sewage samples per day during peak times, the government worked closely with industry partners to decentralise the testing works to multiple commercial laboratories and ensure sample delivery with a well-established sample labelling and logistics system.

The in-manhole auto-sampler

Synergy between surveillance and intervention

Throughout the entire pandemic, the team was working round the clock with sewage sampling starting at 06:00 hrs and the corresponding testing and data analysis reports being produced before the next dawn. This quick response time could only be made possible through the team’s creativity and new inventions, including the use of innovative in-manhole auto-samplers to facilitate sewage sampling during morning rush hours in the busy city and the development of a real-time, interactive Web-GIS platform to enhance the efficiency of real-time data handling, analysis, and sharing among government departments. All this new research was brought together to become a powerful tool for the government to fight the virus in real-time, enabling the planning and execution of immediate responses such as intervention measures and relaxation and tightening of social distancing measures.

All this scientific data formed a quantitative basis for the government to devise appropriate anti-epidemic measures at district levels. For example, a Compulsory Testing Notice (CTN) or Restriction-Testing Declaration (RTD) would be issued under the Prevention and Control of Disease Ordinance (Cap.599) when surveillance results showed a positive and significant COVID-19 signal in a certain district. This mandated all residents or individuals who were present in the area to carry out PCR tests within the next few days, or the introduction of restricted movement and intensive PCR testing of all residents in a building or street block during the hours of 18:00-06:00 to allow healthy residents to go to work the next morning.

During the course of only a few months in the fifth wave of COVID-19, the government mobilised thousands of staff and contractors to implement 306 RTD operations at high-risk residential buildings and at street-block level based on the testing results. Over 26,700 confirmed cases were detected with the majority being asymptomatic (i.e., cases still potentially spreading the disease in the community if left undiagnosed).

Cumulative cases estimated from Viral Concentration

In parallel, more than 100,000 Rapid Antigen Test (RAT) kits were distributed every day to areas of moderate risk to encourage local residents, cleansing workers and security guards of property management companies to conduct self-testing. If tested positive, home quarantine orders would be issued to further reduce the chance of transmission. These measures formulated a tiered and cost-effective approach to combat the disease in the local setting, taking the risk factor, in the form of spatial viral concentration, into full consideration. With this sewage surveillance in place, the community was alerted to the risk of infection, and the spread of the virus was contained without resorting to extreme measures like city lockdowns or compulsory territory-wide PCR testing.

Use of forecast models to predict the epidemic trend

Upon collection of such an extensive amount of sewage data, the team also successfully developed two real-time forecast statistical models to predict case counts. Through detailed data analysis, individual models were developed for case counts reported by positive PCR (through community clinical testing) and RAT results (through self-testing). The models were simple to use with only per capita sewage viral concentrations as inputs, with a lead time of two to four days sufficient for the government to adequately plan for intervention measures. The success of these two models demonstrated the feasibility of real-time epidemic tracking even in a densely populated city like Hong Kong with a complicated sewerage system. The number of new cases detected by PCR had a four-day lag time to the sewage signal, while the number of new cases detected by RAT had a two-day lag time to sewage signal.

The two models were not only used for predictions of new cases but also used to estimate the doubling time/half-life of the spread of the disease in the community. For example, the per capita viral concentration doubling time for 1-13 February and 13-27 February were two days and five days, while the half-life for 27 February-11 April was seven days. The information was then transformed to predict the number of new cases for up to a month, as shown in Figure 3. This provided very important support to enable the government to determine if there was any need to change intervention measures.

Intervention actions due to positive sewage results

These two models were subsequently verified and used to hindcast the cumulative number of COVID-19 infected cases in Hong Kong from the beginning of the fifth wave. The models estimated that a total of 6.5 million cases had actually occurred within a period of a year, an amount accounting for 85% of the population in Hong Kong. The results also revealed that the city could have already built up a strong community immunity against the disease, providing evidence which led to the government relaxing most of the anti-epidemic measures in early 2023.

As I commented at the recent IWA LET conference: “The interactive sewage surveillance programme serves as the headlights when driving in the dark. It provided valuable information to enable us to decide whether we should tighten or relax our intervention measures. Furthermore, an accurate estimation of the number of new cases during the pandemic was as important as the prediction of the trend of disease spreading. Mutation can result in a more severe or a milder virus and is largely dependent on the selection pressure, such as vaccination, compulsory testing, quarantine, community immunity and social distancing. We need time for the community to develop immunity and at the same time we need to ensure the health system will not be overloaded. By monitoring the per capita sewage viral concentration and comparing it with the number of the severe and fatal cases through the two models developed, we could estimate the spread and the severity of the disease at different times, which helped us to understand whether the control strategies were effective or not and whether there was a need to tighten or relax the intervention measures.”

Using this scientifically proven sewage surveillance technology as a robust instrument for policy formulation, the HKSAR government has innovated to enhance the conventional public health decision-making process by introducing science to inform important public health choices through a risk-based approach. With this well-established and proven system now developed and verified through real-life application, it can be further utilised as an early warning system, alongside other data sources, to detect possible re-emergence of COVID-19, as well as outbreaks of other sewage-borne diseases to afford early warning and pre-empt potential overloading of clinical capacity, especially in less developed regions. To better prepare for future potential epidemic situations, Hong Kong has also expanded the sewage sampling network to cover 7000 sampling sites across the territory. l


The Environment and Ecology Bureau and the Environmental Protection Department would like to extend their profound appreciation to collaborators including the University of Hong Kong for its sewage surveillance technology research and development; the Chinese Academy of Sciences for providing professional advice on testing methodology and data analysis; the Drainage Services Department and the Department of Health of the HKSAR government for providing continued support in sampling and serving as the epidemiology advisor of the programme; the CMA Industrial Development Foundation, ALS Technichem (HK) Pty and the Castco Testing Centre for providing sewage testing services; Hong Kong University of Science and Technology for conducting the inter-laboratory comparison programme; and all the Hong Kong civil servants who contributed to the fight against COVID-19.

More information

‘The city-wide full-scale interactive application of sewage surveillance programme for assisting real-time COVID-19 pandemic control – A case study in Hong Kong’, Science of The Total Environment 875 (2023): 162661.

The author

Dr. Samuel CHUI at IWA LET May 2023

Dr Samuel HK Chui is Director of Environmental Protection, Government of Hong Kong SAR, China. He gave a keynote speech on this article’s theme at IWA’s 18th Leading Edge Conference on Water and Wastewater Technologies (LET2023) in Daegu, Korea, on 30 May.

Key timeline

In December 2020 the first sewage Compulsory Testing Notice (CTN) was issued and nine asymptomatic cases were identified within two weeks.

Virus persists in Hong Kong sewage after mass tests

In January 2021 the HKSAR government sealed off an area of high-density population and old buildings for the first sewage Restriction-Testing Declaration (RTD).

In June 2021 sewage tests identified the first Delta variant case three days before medical confirmation.

From January 2022 sewage surveillance was extensively used to issue CTN and RTD and distribute Rapid Antigen Test (RAT) kits.