Combining Rapid Antigen Testing and Syndromic Surveillance Improves Community-Based COVID-19 Detection in Low-to-Middle-Income Countries


 Diagnostics for COVID-19 detection are limited in many settings. Syndromic surveillance is often the only means to identify cases, but lacks specificity. Rapid antigen testing is inexpensive and easy-to-deploy but concerns remain about sensitivity. We examine how combining these approaches can improve surveillance for guiding interventions in low-income communities in Dhaka, Bangladesh. Rapid-antigen-tests and PCR validation was performed on 1172 symptomatically-identified individuals at home. Statistical models were fit to predict PCR status using rapid-antigen-test results, syndromic data, and their combination. Model predictive and classification performance was examined under contrasting epidemiological scenarios to evaluate their potential for improving diagnoses. Models combining rapid-antigen-test and syndromic data yielded equal-to-better performance to rapid-antigen-test-only models across all scenarios. These results show that drawing on complementary strengths across two rapid diagnostics, improves COVID-19 detection, and reduces false-positive and -negative diagnoses to match local requirements; improvements achievable without additional expense, or changes for patients or practitioners.

can match PCR in terms of both sensitivity and specificity. However, lower 98 sensitivity and specificity may be admissible depending on the scale and im-99 pact of misclassification. [17] Low specificity means more common COVID-19 100 misdiagnoses (false positives), leading to unnecessary self-isolation, which 101 is expensive to individuals and society. [18] Low sensitivity means COVID-102 19 cases will be missed (false negatives) and mitigation measures not put in 103 place. [19] These misclassifications are complementary for a given diagnostic, 104 meaning increasing specificity will lead to decreased sensitivity, and vice versa. 105 The typical approach is to maximise the number of correct classifications 106 and assume that both misclassification types are equally costly. But, if the 107 disease is prevalent or increasing, false negatives will have an outsized and 108 costly impact. [19] Or, under low prevalence, false negatives will be correspond-109 ingly low so even a high false negative rate (low sensitivity) will have modest 110 impact, but small decreases in specificity will lead to a large number of expen-111 sive false positives. [20] In practice the situation will be more nuanced and mod-112 ulated by testing capacity constraints, requiring a balance to be struck. [17] 113 The best diagnostic approach for surveillance will therefore be one where timised combined models achieve equal-to-much-lower error rates than the 125 next best method in all metrics, and how integrating data from multiple rapid 126 testing methods can improve diagnostics, particularly when adapted to local 127 situations.

129
Of 1241 participants enrolled by community support teams across Dhaka, 130 1172 had complete data available for analyses. The remainder were removed 131 due to duplicated sample identification codes that prevented reliable matching 132 of test results to symptom metadata. These duplications occur at random, 133 due to human error, and we do not believe they could bias results. Patient 134 summaries by age, gender, case positivity and symptoms are presented in Ta-135 ble 1. Case positivity in Dhaka increased from 15.8% to 23.8% from the first 136 (19th-26th May) to the last week (4th-11th July) of the study, corresponding 137 to prevalence rising from 1.4 to 13.8 confirmed cases per 100 000 people.
138 Table 1: Breakdown of patient numbers by age and gender, in relation to case positivity by PCR and reported symptoms (both as % rounded to nearest integer). Although age is binned here, raw age in years was used for analyses. Furthermore, in the survey non-binary genders were permitted but none reported.

161
Scenario-specific classification performance is shown in Figure 3. Across all 162 scenarios (defined in Table 2), the best models in Class 3 that used both 163 the rapid antigen testing and syndromic data performed equally well or bet-164 ter than the other two model classes. In Scenario 1 ("Agnostic"), models in  In Scenario 3 ("Low Incidence"), Model Class 2 (syndromic data) again per-173 formed worst, and Model Class 3 achieved the lowest error, with Model Class 174 1 falling between the two (closer to Class 3 than 2). The error in Class 1 was 175 0.02 and the median errors ranged from 0.19 to 0.2 for Class 2, and 0.18 to 176 0.2 for Class 3 (Figure 3).

177
The candidate models are chosen as a result of a selection process and per-178 formed much better than more complex models (i.e. with 5 or more symp-  Figure 2: Interquartile ranges for receiver operating characteristics (ROC) for rapid-antigentesting-only approach (Model Class 1) and posterior median and interquartile range ROC for Class 2 (syndromic-data only) and 3 (syndromic and rapid-antigen-testing data) models. Note that in Model Class 1 the ROC is a single value as the binary test has a single sensitivity and specificity. In Models Class 2 and 3, the ROC are curves which demonstrate the performance of the model for any hypothetical scenario as defined by the axes (as opposed to Figure 5 which demonstrates model performance in specific epidemiological scenarios which are realisations of single points in this space). improvement is seen under the scenario of "Epidemic Growth" (see Table 2)   Figure 4: Schematic description of identification of likely COVID-19 cases by community support teams (CSTs) and model definitions. CSTs collect syndromic data (age, gender and presence/absence of 14 predetermined symptoms), and two sets of naso-pharyngeal swabs (for rapid antigen testing and PCR). We used three model classes: rapid-antigen-test-only in 1, syndromic data only in 2, and both rapid-antigen-test and syndromic data in 3. The PCR result is used to train and test each model using temporal cross-validation.
over models which implicitly assume independence between symptoms. By test-positive and -negative, allowing the model to adapt solely to the latter.

309
The models were fitted to the data using Bayesian inference techniques based To make exploring these models computationally feasible and to reduce the risk of overfitting, we carried out two rounds of model selection. A subset of symptoms are identified using relationship between each symptom and PCR-status identified by the corresponding model. From this subset of symptoms, a more exhaustive search of potential models is then conducted to identify the best symptomcovariate relationships, using temporal-cross validation to measure model performance. The best model for each level of complexity (i.e. number of symptoms) are then used as our candidate models. Only these final models are used for classification. This reduces the set of models tested as classifiers from >131 000 to just four per model class.
reduce the risk of overfitting. The large number of symptoms corresponds 319 to many potential model configurations (>131 000 for 14 symptoms and two 320 covariates) which might perform well on the test sets by chance (even under 321 temporal cross-validation) but lack transferability to novel situations. By  Table 2: Requirements and performance criteria for each epidemiological scenario. The requirement refers to a base level of performance the model must achieve. These requirements were determined through discussion amongst the authors and colleagues at IEDCR. The performance criterion is used to determine which model performs the 'best' given that the requirement has been met. We assessed models using three sets of increasingly policy-relevant criteria.  Table 2).

336
We scored the models' predictive power using cross-entropy (defined in Sup-  We compare model performance under three scenarios (using error terms de-355 15 scribed in Table 2) developed for illustrative purposes through discussion with 356 colleagues at IEDCR. In Scenario 1, we do not consider epidemiological con- The study funders did not participate in the design of the study, data collec-369 tion/analysis/interpretation, or in the writing of this paper. The data and statistical code used in this study are available in a GitHub 385 repository at https://github.com/fergusjchadwick/COVID19_SyndromicRAT_public. We examined the ability of the two imperfect identification methods, syn- correlations are more familiar. We thus formulate the multivariate probit as:

Declaration of Interests
Model Class 3 combines the two data sources. We utilise the specificity of shrinkage. We used minimal shrinkage, η = 1) [28] . More informative priors 510 that incorporate spatio-temporal effects, for instance, would be natural exten-

539
By using general predictive power to narrow down the number of candidate 540 models and then testing those models, we are more likely to choose models 541 that generalise well to new data. It was clear when fitting the models that 542 there were "jumps" in performance (as defined below) between models con-543 taining five and four symptoms, so the models with one to four symptoms 544 were used as the candidate models. Zero symptom models were not included 545 in the analysis as they do not correspond to a feasible policy (with covariates 546 they would require governments to ask individuals of a given gender and age 547 as COVID-19 positive, and without covariates they would involve randomly 548 assigning individuals as COVID-19 positive). vector of probabilistic predictions (p(y) ∈ (0, 1)) as follows: The resulting score is comparable across all methods for assigning predictions 559 where the same test data are used, allowing us to compare predictions from to generate a ROC curve. In our epidemiological scenarios (outlined below) 576 we use our ROC curve calculations to identify single thresholds which yield a 577 required error rate. 578 We strongly emphasise that generic performance here is only used to show  Table 2). We identify 583 the threshold,p, at which the requirement is most closely exceeded (i.e. if 584 the requirement is an error rate should be a maximum 15%, the threshold 585 that produces an error rate below 15% but as close to 15% as possible will be 586 chosen).

587
In Scenario 1, we do not consider epidemiological context but simply minimise words, the arithmetic mean could be large because it has a very high TPR but 595 a small TNR, whereas the harmonic mean will maximise both TPR and TNR.

596
While conceptually the harmonic mean is better suited than the arithmetic for 597 this use case, both produce qualitatively the same results for these data.

598
Scenario 2 corresponds to the situation in Bangladesh at time of writing 599 (September 2021), with COVID-19 cases beginning to rapidly increase again.

600
Under these circumstances, false negatives are extremely costly relative to 601 false positives due to the exponential growth of the disease.

602
In Scenario 3, the pandemic is not declining but maintaining a steady rate 603 of cases. In this situation, policy-makers may be keen to keep false positive 604 diagnoses low to prevent lockdown fatigue and to keep the workforce active. Once the PVF has been screened, the result can be either of these three:

Verified Virus Fighter (VVF):
A PVF who has been screened and has high fever and with relevant signs/symptoms of respiratory disease (for example cough, shortness of breath (in last 15 days), sore throat) or the loss of the sense of smell.
OR an individual who tested COVID-19 positive in the last few days.

PVF with follow-up:
A PVF who has been screened and whose body temperature is between 99.0oF to 99.4oF AND who has at least one sign/symptom of respiratory disease (for example cough, shortness of breath (in last 15 days), sore throat, the loss of the sense of smell).

Cleared Virus Fighters (CVFs):
A PVF whose body temperature is below 99oF or who does not exhibit any symptoms of respiratory disease (for example cough, shortness of breath, sore throat, the loss of the sense of smell).

Vulnerable Individuals:
Certain individuals are at higher risk of developing complications and dying from COVID-19, these include older individuals (50 years or older in the context of Bangladesh), diabetics, hypertensive individuals, individuals with respiratory diseases such as COPD or those with compromised immune systems. Pregnant women are also a high-risk group for COVID-19 related adverse outcomes.

Scope
For use by CSTs, AMS/VAMs, telemedicine doctors, field Implementation teams and their support teams operating in urban & rural areas of Bangladesh to carry out surveillance for COVID19 and identification and protection of vulnerable groups.

Purpose
The purpose of this Standard Operating Procedure (SOP) is to provide a brief overview of the workflow of CSTs work to a) identify vulnerable individuals efficiently, b) identify the PVF c) take the necessary steps to follow once someone identified as VVF, CVF or follow up PVF. The SOP will link to other technical SOPs and provide guidance to the following activities. 6. Provide SRHR telemedicine numbers to all females aged 15-49 years in the household.

Procedure:
Once the CSTs are trained and grouped into teams and assigned to a particular ward, they will need to do the following coordination activities:

A. Coordination with local authorities Urban:
The AM/VAM (with support from the Field Implementation (FI) team of BRAC will organize inperson meetings for the different wards and zones of the city corporation. The participants should include: the focal person from the ward councilor, the ward councilor, and the zonal executive officer (ZEOs) and the Deputy Chief Health Officer (DCHO). The AM/VAM will support the FI team to inform the local police station about the CSTs working under their jurisdiction. This will include sharing a list of each CST member (along with their photos) working in their particular wards.
Steps in organizing the meeting: a) They will be provided the contact details of focal persons and members from Ward councillors. b) All physical distancing rules have to be followed: the meeting will only include essential individuals to prevent overcrowding. There should be a minimum of 1-metre distance between each individual. c) All participants will perform hand hygiene on arrival and when leaving the meeting and they will all wear masks d) Prior to the commencement of the meeting, the meeting venue, including chairs and tables, will be cleaned with disinfectants, especially hard surfaces. Follow-up: On a regular basis, the CST should share activity updates with the local authority /focal person through telephone. The AMs/VAMs will update the FI team regularly, who will also facilitate discussion with the ward councilors/ZEOs/DCHO.

Follow-up:
On a regular basis, share activity updates with the local authority through telephone.
Note: Representatives from the different partner organizations will try to attend the field coordination meetings.

B. Maintenance/Handling of logistics
Refer to SOPS 3 and 5 for materials needed for CSTs to carry out their duties in a safe and professional manner.
CST members will be provided with an Infrared thermometer, oximeter, Wrist watch BP monitor, three-layer cloth mask, measuring tape, gloves, goggles, bleaching powder, disinfectant containing 70% alcohol, id card and vest. They can keep the logistics in house of one of the members.
Infrared thermometer, oximeter, three-layer cloth mask will be provided in the training by FAO.
Other logistics like more masks and sanitizers will be provided and managed by BRAC.

Word-of-mouth:
The CSTs will aim to visit an agreed number of households (but the focus should be on complete and comprehensive screening -it is more important to identify VVFs and vulnerable than to maximize household visit numbers). During the visit they will identify PVFs by word of mouth.

Government Hotline:
• The CSTs will also have to visit households with PVFs identified through the government hotline. The AM/VAM will contact the PVFs who called in Government hotlines in the last 2-3 days using the phone number used in the call. They will communicate with individuals, guided by a talking point tree, which explains the CST activities and requests the person to allow a screening visit from the CSTs. The contact information will be passed to the specific CST team through the CST mobile app.
• The CST will receive a list of phone numbers of PVFs in their CST mobile app. • One of the CST members will call the PVF, introduce themselves, describe the purpose of visiting their house and request to schedule a visit. (They will follow the leaflet on FAQ "Coronavirus and CST team related information).
• If the PVF is reluctant for a visit, the CST member will try to convince them using interpersonal communication skills. If they still do not agree for the household visit, then the CST member will try to advise them about quarantine over the phone and also ask if they require food or medicine support. The CST will also check if there is any pregnant woman in the house, or any woman who has given birth in the last 6 weeks.

Note:
The CST (AM/VAM) should make every effort to contact the PVF, this includes calling each number three times before giving up if it is not answered. If the individual is not willing to have a home visit, the CST will try to counsel them; this may include two calls to try to arrange a home visit.
D. The total Household Visit will include three major activities: 5. If the household member doesn't agree then the CST will continue to finalize the form without collecting any phone number and name, give them the CST sticker, thank them and leave the house.
6. If the household member agrees to continue, the CST will note down then his/her name and phone number as the primary respondent and proceed with the rest of the household form questions (basic questions about vulnerability: Age breakdown of HH members (most important to obtain accurate information about number of individuals over 50 years of age); pregnant, hypertension, diabetes and COPD).
PVF Screening: 7. The CST will then screen each member with COVID-19 symptoms with the PVF screening form will need to be completed for each PVF.
8. At first, the CST will seek consent for PVF screening, if the person doesn't agree then the CST will end the PVF screening. If the person agrees, then the CST will continue with the PVF screening form. This PVF screening process will need to be repeated for each household member with COVID-19 symptoms. 2. The CST will explain why it is necessary to identify vulnerable people.
3. The CST will follow the App to fill up the first part of the health screening form.
4. Depending on whether the person has one or more declared health issues, the CST will seek consent to do a physical examination to help identify undiagnosed conditions (for example, if an individual says no or does not know about having high blood pressure, the CST will measure their blood pressure).

Consent Statement:Do you want to have a physical examination now, this will include you
measuring your own waist and might include measuring blood pressure.
5. Once the person agreed to the screening process, please follow SOP 4 c_on how to measure BP and SOP 4 d_on how to measure waist circumference. 6. The CST will proceed to ask remaining questions on the screening form.
7. At the end of the screening form, the CST should ask for the respondent's name and phone number.

C. Steps to take for individuals identified as vulnerable:
For individual identified as Elderly, Diabetic, Hypertensive and COPD-provide counselling based on SOP 09 and refer them to CST telemedicine.
Pregnant women: Refer to SRHR telemedicine D. Steps to take after PVF screening based on the screening result

Steps to be taken if PVF is identified as VVF:
• Please refer to SOP 5 Home Family Quarantine Support". Support the VVF in maintaining quarantine for 14-days along with their entire households.
• Measure oxygen saturation levels and enter the level into the App. Take appropriate action based on the oxygen saturation levels.
• Connecting to the telemedicine doctor who will determine the severity of the VVF's symptoms. Depending on the severity, the doctor will recommend a course of action.
(Refer to SOP 7: Dedicated Medical Guidance Call Centre for VVF). If it is not possible to connect to the doctor, the CST can leave the number with the VVF to call later.When connecting the VVF with a Telemedicine doctor CST needs to inform the doctor about this.
Doctor will decide the severity of the case.
• Provide essential medicine and food support for low income households or arrange for these things to be procured by friends or neighbors.
a. The CST will teach the other members of the household on how to avoid direct contact with VVFs while still supporting and motivating the VVF fight against COVID-19. The CST will ensure that the neighbors will understand the role of the VVF in the fight against COVID 19 and are ready to help them. b. The CST will proceed to include information on all household members as per the app specifications (name, phone number, age, gender and relationship with the VVF). If another household member is showing COVID-19 symptoms and wants to be screened as a PVF, only then the CST will screen them as a PVF, otherwise only the information mentioned above needs to be collected for each household member.
c. Carry out scheduled follow up visits to ensure adherence to proper quarantine, check if medicines or food is needed and to check if symptoms have worsened.
• At the 14th day of quarantine period, doctors from telemedicine will call VVF to find out the current status if no further sickness is in the household, they are all free to end isolation.

Steps to be taken if the PVF is identified as PVF with follow up
The CST will counsel them about monitoring symptoms very closely, and call the CST right away if l if the symptoms worsen a. The CST will share their phone numbers if they need further support and will advise of any follow up visits b. In any event, the CSTs will revisit him/her within two days to reassess their symptoms and start the whole screening processes again by following the relevant section in the App.
c. The App will determine if the person is VVF, PVF with follow up or cleared PVF; d. If the app changes the status of the PVF with Follow-up to VVF, then the CST will follow SOP 5 as outlined above.
e. If the app keeps the status of the PVF with follow up, the CST will ask the PVF to contact them if the symptoms worsen. If PVF doesn't contact the CST, CST does not need to visit the household further.
f. Cleared PVF, the CST will follow the SOP 5 for these categories.
g. The CST will advise the PVF with follow-up to call the CST immediately if symptoms worsen. The CST will always also advise the entire household to wear masks when going outside their homes and to request visitors to wear masks when visiting. h. If anyone in the family develops cough or fever, they may report again contacting their local CST or using 333 or 16263.

Monitoring VVFs and PVFs with follow-up:
• Monitoring visits will clearly schedule and are designed to:

Standard Operating Procedure (SOP) 3 for Screening Potential Virus Fighters Preamble
Once a PVF is identified and details entered into the CST Mobile App, the job of the CST is to screen the CST for COVID19 as soon as possible. The screening process is assisted by the App, which will confirm the status of the PVF. Once the PVF has been screened, the result can be either of these three:

Verified Virus Fighter (VVF):
A PVF who has been screened and has high fever and with relevant signs/symptoms of respiratory disease (for example cough, shortness of breath (in last 15 days), sore throat) or the loss of the sense of smell.
OR an individual who tested COVID-19 positive in the last few days.

PVF with follow-up:
A PVF who has been screened and whose body temperature is between 99.0 o F to 99.4 o F AND who has at least one sign/symptom of respiratory disease (for example cough, shortness of breath (in last 15 days), sore throat, the loss of the sense of smell).

Cleared Virus Fighters (CVFs):
A PVF whose body temperature is below 99 o F or who does not exhibit any symptoms of respiratory disease (for example cough, shortness of breath, sore throat, the loss of the sense of smell).

Vulnerable Individuals:
Certain individuals are at higher risk of developing complications and dying from COVID-19, these include older individuals (50 years or older in the context of Bangladesh), diabetics, hypertensive individuals, individuals with respiratory diseases such as COPD or those with compromised immune systems. Pregnant women are also a high-risk group for COVID-19 related adverse outcomes. Community Support Team (CST): In urban and residential areas, the CST will consist of at least two volunteers from different volunteer organizations (e.g., Platform, CDP, Utshorgo foundation, Young Bangla), students from the communities and/or volunteers nominated by the Ward councilors.
In urban slum area the CST will consists of two Shasthyo Kormi (SK) from BRAC.
Each CST team will be assigned to one ward, and they will be supervised by Area Managers (AM) or Volunteer Area Managers (VAM).
These AMs/VAMs will be responsible for multiple wards (and hence multiple CSTs).

Scope
The SOP is used to determine the status of PVFs by identifying signs and symptoms of COVID19.
The PVFs may be identified by the hotline, the community, or as close contacts of a VVF or identified COVID19 patients.

Purpose
The purpose of this SOP is to provide detailed guidelines to the CST on how to screen PVFs) to determine if they are a) VVFs b) PVF with follow-up c) Cleared PVF

Logistics required (in necessary quantity as per visit plan) for CST:
• Smart phone/tab • All necessary supportive medicines (first line treatment advised by government telemedicine number 16263) will be carried for distribution to VVF household.
• Phone numbers of local ME Procedure 1. The CSTs will make a daily plan for household (Khana) visits to identify PVF by word of mouth as part of their daily work and the PVFs on the App provided by Area managers.
2. The CST will visit household as per given daily target and look for PVFs to be screened.
• CST will start the conversation with permission and introduce themselves, explain the HH members why they are here.
• Then they will ask if someone in the household is sick or want to be screened as a

PVF
• If there is any sick person or the member of household want to be screened, please follow the steps from "6 to 14" • If the household denied for screening please follow steps "4 and 5".
3. The CST calls the number from the app to confirm the name and number of the PVF and to request a screening visit. If the number does not answer, they should try three times before reporting the number as not answering.
4. If the PVF is reluctant for a visit, the CST member will try to convince them using good interpersonal communication skills. If they still do not agree for the household visit/screening, then the CST member will try to advise them about quarantine over the phone and also ask if they require food or medicine support. The CST will also check if there is any pregnant woman in the house, or any woman who has given birth in the last 6 weeks.
The CST should not give up on the home visit but call again at another time.
5. The CST should use good communication skills to build trust with the PVF and household members before starting the screening. They should remember to treat the PVF as an equal and to respect his/her concerns. They should explain clearly why they are wearing PPE and why they will be taking measurements and asking questions. They cannot enter the house and commence the screening without permission of the household members. 10. Based on the signs and symptoms and temperature reading, the mobile app will determine if the PVF meets the definition for a VVF or PVF with follow-up or a Cleared PVF.
11. The CST will also check if there is any pregnant woman, any woman who has given birth in the last 6 weeks (42 days) or any vulnerable people in the household.
12. If the PVF is a VVF, then the CST will take the following steps: a) The CST will measure the blood oxygen saturation of the VVF using the pulse oximeter and record the reading in the CST mobile app (see SOP 4b. Using the oximeter). b) If the oxygen saturation level is equal to or below 93%, the CST should explain to the VVF that he needs specialized medical treatment and immediately call the AM/VAM for assistance to take the VVF to hospital.
c) The CST should proceed to ask the rest of the questions as prompted in the CST mobile app.
c) The CST should add information (name, age, sex, telephone number and relationship to the VVF) of each household member in the CST mobile app.
d) These household members should be screened as PVFS.
e) The CST will check if any member of the household (including the PVF who was just screened) is either pregnant or a breastfeeding. The household will be given the OGSB number to call for any advice on referral to a hospital or any other issue. If a female household member is pregnant: The CST will advise them to go to a health facility for regular antenatal visit, and to deliver in their facilities.
-If a female household member is a breastfeeding mother: The CST will advise them to wear masks while breastfeeding, and for them to consider family planning.
f) The CST should connect to the telemedicine doctor and hold the conversation on speaker phone so that both the VVF and CST can hear. The medical expert who will determine the severity of the VVF's symptoms and depending on the severity, will recommend a course of action. (Refer to SOP 9: Dedicated Medical Guidance Call Centre for VVF). If it is not possible to connect to the doctor, the CST can leave the number with the VVF to call later.
g) The CST should then follow the Home Family Quarantine Support SOP 5 for guidance on counselling the VVF and their family for maintaining 14-day home quarantine, implementing IPC within the household and support measures.
Version 08 h) The VVF should be advised that there will be personal follow up visits on days 3 and 7 and then a phone check up on day 10.
i) The CST will then ask the VVF if they want to be tested for COVID-19, if the opportunity arises.
Before recording the answer in the CST mobile app, the CST will clearly describe the consequences of agreeing to be tested as described below: Once the PVF has been screened, the result can be either of these three:

Verified Virus Fighter (VVF):
A PVF who has been screened and has high fever and with relevant signs/symptoms of respiratory disease (for example cough, shortness of breath (in last 15 days), sore throat) or the loss of the sense of smell.
OR an individual who tested COVID-19 positive in the last few days.

PVF with follow-up:
A PVF who has been screened and whose body temperature is between 99.0 o F to 99.4 o F AND who has at least one sign/symptom of respiratory disease (for example cough, shortness of breath (in last 15 days), sore throat, the loss of the sense of smell).

Cleared Virus Fighters (CVFs):
A PVF whose body temperature is below 99 o F or who does not exhibit any symptoms of respiratory disease (for example cough, shortness of breath, sore throat, the loss of the sense of smell). Certain individuals are at higher risk of developing complications and dying from COVID-19, these include older individuals (50 years or older in the context of Bangladesh), diabetics, hypertensive individuals, individuals with respiratory diseases such as COPD or those with compromised immune systems. Pregnant women are also a high-risk group for COVID-19 related adverse outcomes.

Scope
For use by Community Support Teams (CSTs) once a PVF has been declared a , VVF and recommended to follow 14 days of isolation.

Purpose
The purpose of this SOP is to provide guidance for CST on how to advise and support VVFs on self-isolation and on quarantine of household contacts.
The steps in screening of the PVF is covered in the SOP 3 Quarantine Screening.

Procedure
1. Once the PVF has been confirmed to be a VVF by the CST using the Mobile App, the CST should advise the VVF and family of his/her status and explain the role of the VVF in controlling the spread of COVID19. It is crucial that the CST explain the importance of his/her actions for the community and Bangladesh and gain the agreement of the VVFs and their families. Good communication skills are needed. Key points to be made include: • The COVID19 virus is very contagious and can be spread through sneezing and coughing and touching contaminated surfaces. But the virus can easily be killed by cleaning and disinfection.
• Most people do not get very sick, but a small group may need to go to hospital.
• By isolating the VVF is preventing spread of the virus and is working to protect his/her community. If the virus spreads uncontrollably, the hospitals will be unable to cope and many people will die (can use the fish pond example).
• Isolation is a selfless act that helps others; the VVF is a hero.
• After 14 days of isolation the VVF should be over the COVID virus; it will also be clear if the other family members have also caught COVID19.
• The CST will support the family through the isolation period. c. They should not go out to buy food or collect medicine: they can ask the CST to support them in the process, ask someone else to drop off medicine or groceries at their home or order them by phone or online.
d. They should not allow any visitors, other than the CST or medical persons, in their home.

Household hygiene
1. It is very important to protect other household members from COVID19. The VVF should strictly adhere to the following to prevent infecting other household members: 2.
The VVF should remain isolated in a separate room and stay 1 meter (3 feet) from other members of the house.

3.
He/she and must wear mask and also all family members must wear mask when more than one person is in a room.

4.
Enough food and drinks should be prepared and delivered to the VVF's room but not handed to Him/her. Empty plates and cups should be placed into a bucket at the end of each meal and removed and washed in hot soapy water.

5.
If the family uses common bathroom, specific bathroom times should be scheduled for the VVF and space given for him/her to move to the bathroom and back to the bedroom. 6. The family members should continue to communicate with the VVF and provide them with company and reassurance from a distance of 3 feet while wearing masks.

7.
The VVF should not share dishes, drinking glasses, cups, eating utensils, bed linen, clothes or towels with the rest of the family.

8.
VVFs who are breastfeeding mothers can breastfeed their infants wearing a mask. They should thoroughly wash their hands with soap and water or sanitize their hands with alcohol-based hand rub before breastfeeding.
9. To reduce the spread of infection in the home, the VVF and other household members should do the following: a. Wash their hands with soap and water often, for at least 20 seconds, or use an alcohol-based hand rub when soap and water is not available. The CST members will show them how to correctly wash their hands and show a sample of alcohol based hand rub.
b. The VVF should wear a cloth mask that covers the nose and mouth when he/she must be around other people or animals, including pets. The mask is not necessary when the VVF is alone.
c. All the household members must wear masks at all times inside the house (except when someone is completely alone).
d. All the household members should sneeze and cough inside the mask; for sudden onset of coughing or sneezing when they are not wearing the mask, they should cover their mouth and nose with a tissue or sleeve (not hands), put used tissues in the covered waste bin immediately and wash hands afterwards. If the mask gets soiled by cough or becomes wet it should be changed.
e. The responsibility of taking care of VVF should be given to the healthiest family member who is without any comorbidities such as diabetes, hypertension, cancer, heart disease, chronic respiratory disease.
f. Surfaces that are touched often (like door handles, bathrooms, kettles, light switches, chair arms) should be cleaned regularly using household cleaning products and disinfected with 0.2% bleach.. Electronic items such as phones should be cleaned with alcohol. g. If a caregiver or other person needs to clean and disinfect a sick person's bedroom or bathroom, they should wear a mask and disposable gloves prior to cleaning. They should wait as long as possible after the VVF has used the bathroom before coming in to clean or use the bathroom. The area should be cleaned first with soap and water followed by disinfection with 0.2% bleach.
The disinfected area should remain wet for 5 minutes and then excess bleach cleaned up with a clean cloth.
h. Wash the cloth mask with warm water and detergent every day or soak in 0.2% dilute bleach (Add 2 teaspoons of bleaching powder to 1 liter of water) for at least 1 minute, rinse with water, and then let air dry in sunlight if possible.
i. The house should be cleaned with normal household products, such as water and detergent, followed by a 0.2% disinfectant bleach. CST members will demonstrate how to make a 0.2% bleach solution (see below) and give each family a 250 mg packet of powder bleach to clean the surrounding surfaces around the VVF and the bathroom after use.
j. Used tissues and disposable cleaning cloths should be placed in garbage/ polythene bags and then put into a second bag and tied securely. The bag should be stored for 3 days before putting it in the outside bin. Other household waste may be disposed of as normal.
k. Laundry should be washed in the usual way. Laundry that has been in contact with an ill person can be washed with other people's items but they should not be shaken as this may spread the virus in the air.

10.
To stay well while at home, the VVF and any ill household members should: a. drink plenty of water to stay hydrated,

11.
If the household includes a vulnerable individual (someone who is 60 years old or over, has a long-term condition, is pregnant or has a weakened immune system), the household should try to move him/her to another house for 14 days.

12.
If the vulnerable person must stay in the home, the VVF and the vulnerable individual should try to keep away from each other as much as possible by: 13. If the VVF or another ill household member needs medical help during Family Quarantine/Isolation, he/she should not go to a clinic, pharmacy or hospital. He/she should stay at home and call the community support team or contact the telemedicine doctor dedicated to VVFs . The CST will have the number for the local ME. 14.
The CST team members should follow up physically with the VVF and his/her household members on the 3rd and 7th day. During the visit they will check VVF's temperature, oxygen saturation and confirm if the VVF and his family are maintaining quarantine (please see SOP 1 Process Flow_Urban Areas for details) Data on body temperature, oxygen saturation, adherence to home quarantine, will be recorded through the CST mobile app during the follow-up visits. The CSTs will also follow-up through telephone on the 10th day to make sure VVF and his family maintained home quarantine properly and also to enquire if they require further food support/medical attention.

15.
During the follow-up visits, the CSTs should counsel the VVF and the family again on steps 2-11 (to reinstate the importance of maintaining quarantine).

16.
The CST will ensure that the neighbors understand the fight and are ready to help them morally and mentally to boost up VVF and the family.