Summary of the protocol of a target trial estimating differences in dengue transmission risk for individuals residing in Wolbachia release sites vs those in control locations who are not
| Component . | Hypothetical cluster-randomized trial . | Target trial emulation . |
|---|---|---|
| Eligibility | Patients who report for undifferentiated febrile illness to any general practitioner clinic, polyclinic or public/private hospital during trial duration and have home addresses that are in trial locations between EW1 2019 and EW26 2022 | Same as hypothetical trial |
| Intervention strategies | Release of wAlB-SG infected male Ae. aegypti in home address of participants. An individual is considered directly intervened by Wolbachia if home address resides in an area that has sustained releases for more than:
| Same as hypothetical trial |
| Intervention assignment | Townships at historically high risk of dengue transmission pre-selected to either receive intervention or control randomly Townships undergo a constrained randomization strategy for control and intervention to prevent chance imbalance and have same historical dengue incidence rates in the pre-intervention period Townships are the unit of randomization | Four long-term Wolbachia field trial townships not randomly pre-selected Control townships undergo a constrained randomization same as the hypothetical trial to match the historical dengue incidence rates of the intervention township in the pre-intervention period Townships are the unit of randomization |
| Release strategy | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week All designated intervention sites would implement intervention at the start of the trial | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week Four long-term Wolbachia field trial townships either adopted an expanding release approach (release sites were gradually expanded to adjacent neighbourhoods) or targeted release approach (focused releases on areas with high Aedes aegypti abundance and persistent dengue transmission) |
| Blinding | Intervention and participants unblinded to intervention | Same as hypothetical trial |
| Follow-up | N/A. Repeated tests were removed | Same as hypothetical trial |
| Enrolment | Unexposed individuals were enrolled if they reported for febrile illness throughout the trial duration and did not reside in an intervention location, whereas exposed individuals were enrolled if they reported for febrile illness and were residing in an intervention location that had interventions for 3, 6 or ≥12 months | Same as hypothetical trial |
| Primary outcome | Test-positive for dengue using highly sensitive and specific laboratory tests14,15 (internally controlled RT-qPCR assay, dengue non-structural protein 1 (NS1) or IgM) | Same as hypothetical trial |
| Causal contrast | Intention-to-treat effect | Observational analogue of the intention-to-treat effect |
| Analysis plan | Intervention (exposure) variable of interest: wAlB-SG infected male Ae. aegypti at home address will be considered as a binary classification based on time since intervention (see intervention strategies) Outcome variable: test-positive status for dengue at point of testing Doubly robust logistic regression will be used to assess the intervention effect by estimating the aggregate odds ratio, which compares the exposure odds among test-positive cases vs test-negative controls. The null hypothesis is that the odds of residing in the intervention locations are the same among test-positive cases as test-negative controls All analyses will be adjusted for environmental and anthropogenic risk factors associated with dengue transmission risk, and the propensity to receive treatment based on the same risk factors. See detailed statistical analysis plan below | Same as hypothetical trial |
| Component . | Hypothetical cluster-randomized trial . | Target trial emulation . |
|---|---|---|
| Eligibility | Patients who report for undifferentiated febrile illness to any general practitioner clinic, polyclinic or public/private hospital during trial duration and have home addresses that are in trial locations between EW1 2019 and EW26 2022 | Same as hypothetical trial |
| Intervention strategies | Release of wAlB-SG infected male Ae. aegypti in home address of participants. An individual is considered directly intervened by Wolbachia if home address resides in an area that has sustained releases for more than:
| Same as hypothetical trial |
| Intervention assignment | Townships at historically high risk of dengue transmission pre-selected to either receive intervention or control randomly Townships undergo a constrained randomization strategy for control and intervention to prevent chance imbalance and have same historical dengue incidence rates in the pre-intervention period Townships are the unit of randomization | Four long-term Wolbachia field trial townships not randomly pre-selected Control townships undergo a constrained randomization same as the hypothetical trial to match the historical dengue incidence rates of the intervention township in the pre-intervention period Townships are the unit of randomization |
| Release strategy | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week All designated intervention sites would implement intervention at the start of the trial | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week Four long-term Wolbachia field trial townships either adopted an expanding release approach (release sites were gradually expanded to adjacent neighbourhoods) or targeted release approach (focused releases on areas with high Aedes aegypti abundance and persistent dengue transmission) |
| Blinding | Intervention and participants unblinded to intervention | Same as hypothetical trial |
| Follow-up | N/A. Repeated tests were removed | Same as hypothetical trial |
| Enrolment | Unexposed individuals were enrolled if they reported for febrile illness throughout the trial duration and did not reside in an intervention location, whereas exposed individuals were enrolled if they reported for febrile illness and were residing in an intervention location that had interventions for 3, 6 or ≥12 months | Same as hypothetical trial |
| Primary outcome | Test-positive for dengue using highly sensitive and specific laboratory tests14,15 (internally controlled RT-qPCR assay, dengue non-structural protein 1 (NS1) or IgM) | Same as hypothetical trial |
| Causal contrast | Intention-to-treat effect | Observational analogue of the intention-to-treat effect |
| Analysis plan | Intervention (exposure) variable of interest: wAlB-SG infected male Ae. aegypti at home address will be considered as a binary classification based on time since intervention (see intervention strategies) Outcome variable: test-positive status for dengue at point of testing Doubly robust logistic regression will be used to assess the intervention effect by estimating the aggregate odds ratio, which compares the exposure odds among test-positive cases vs test-negative controls. The null hypothesis is that the odds of residing in the intervention locations are the same among test-positive cases as test-negative controls All analyses will be adjusted for environmental and anthropogenic risk factors associated with dengue transmission risk, and the propensity to receive treatment based on the same risk factors. See detailed statistical analysis plan below | Same as hypothetical trial |
Summary of the protocol of a target trial estimating differences in dengue transmission risk for individuals residing in Wolbachia release sites vs those in control locations who are not
| Component . | Hypothetical cluster-randomized trial . | Target trial emulation . |
|---|---|---|
| Eligibility | Patients who report for undifferentiated febrile illness to any general practitioner clinic, polyclinic or public/private hospital during trial duration and have home addresses that are in trial locations between EW1 2019 and EW26 2022 | Same as hypothetical trial |
| Intervention strategies | Release of wAlB-SG infected male Ae. aegypti in home address of participants. An individual is considered directly intervened by Wolbachia if home address resides in an area that has sustained releases for more than:
| Same as hypothetical trial |
| Intervention assignment | Townships at historically high risk of dengue transmission pre-selected to either receive intervention or control randomly Townships undergo a constrained randomization strategy for control and intervention to prevent chance imbalance and have same historical dengue incidence rates in the pre-intervention period Townships are the unit of randomization | Four long-term Wolbachia field trial townships not randomly pre-selected Control townships undergo a constrained randomization same as the hypothetical trial to match the historical dengue incidence rates of the intervention township in the pre-intervention period Townships are the unit of randomization |
| Release strategy | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week All designated intervention sites would implement intervention at the start of the trial | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week Four long-term Wolbachia field trial townships either adopted an expanding release approach (release sites were gradually expanded to adjacent neighbourhoods) or targeted release approach (focused releases on areas with high Aedes aegypti abundance and persistent dengue transmission) |
| Blinding | Intervention and participants unblinded to intervention | Same as hypothetical trial |
| Follow-up | N/A. Repeated tests were removed | Same as hypothetical trial |
| Enrolment | Unexposed individuals were enrolled if they reported for febrile illness throughout the trial duration and did not reside in an intervention location, whereas exposed individuals were enrolled if they reported for febrile illness and were residing in an intervention location that had interventions for 3, 6 or ≥12 months | Same as hypothetical trial |
| Primary outcome | Test-positive for dengue using highly sensitive and specific laboratory tests14,15 (internally controlled RT-qPCR assay, dengue non-structural protein 1 (NS1) or IgM) | Same as hypothetical trial |
| Causal contrast | Intention-to-treat effect | Observational analogue of the intention-to-treat effect |
| Analysis plan | Intervention (exposure) variable of interest: wAlB-SG infected male Ae. aegypti at home address will be considered as a binary classification based on time since intervention (see intervention strategies) Outcome variable: test-positive status for dengue at point of testing Doubly robust logistic regression will be used to assess the intervention effect by estimating the aggregate odds ratio, which compares the exposure odds among test-positive cases vs test-negative controls. The null hypothesis is that the odds of residing in the intervention locations are the same among test-positive cases as test-negative controls All analyses will be adjusted for environmental and anthropogenic risk factors associated with dengue transmission risk, and the propensity to receive treatment based on the same risk factors. See detailed statistical analysis plan below | Same as hypothetical trial |
| Component . | Hypothetical cluster-randomized trial . | Target trial emulation . |
|---|---|---|
| Eligibility | Patients who report for undifferentiated febrile illness to any general practitioner clinic, polyclinic or public/private hospital during trial duration and have home addresses that are in trial locations between EW1 2019 and EW26 2022 | Same as hypothetical trial |
| Intervention strategies | Release of wAlB-SG infected male Ae. aegypti in home address of participants. An individual is considered directly intervened by Wolbachia if home address resides in an area that has sustained releases for more than:
| Same as hypothetical trial |
| Intervention assignment | Townships at historically high risk of dengue transmission pre-selected to either receive intervention or control randomly Townships undergo a constrained randomization strategy for control and intervention to prevent chance imbalance and have same historical dengue incidence rates in the pre-intervention period Townships are the unit of randomization | Four long-term Wolbachia field trial townships not randomly pre-selected Control townships undergo a constrained randomization same as the hypothetical trial to match the historical dengue incidence rates of the intervention township in the pre-intervention period Townships are the unit of randomization |
| Release strategy | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week All designated intervention sites would implement intervention at the start of the trial | Standardized release protocol of 1–7 wAlbB-SG males were released per study site resident per week Four long-term Wolbachia field trial townships either adopted an expanding release approach (release sites were gradually expanded to adjacent neighbourhoods) or targeted release approach (focused releases on areas with high Aedes aegypti abundance and persistent dengue transmission) |
| Blinding | Intervention and participants unblinded to intervention | Same as hypothetical trial |
| Follow-up | N/A. Repeated tests were removed | Same as hypothetical trial |
| Enrolment | Unexposed individuals were enrolled if they reported for febrile illness throughout the trial duration and did not reside in an intervention location, whereas exposed individuals were enrolled if they reported for febrile illness and were residing in an intervention location that had interventions for 3, 6 or ≥12 months | Same as hypothetical trial |
| Primary outcome | Test-positive for dengue using highly sensitive and specific laboratory tests14,15 (internally controlled RT-qPCR assay, dengue non-structural protein 1 (NS1) or IgM) | Same as hypothetical trial |
| Causal contrast | Intention-to-treat effect | Observational analogue of the intention-to-treat effect |
| Analysis plan | Intervention (exposure) variable of interest: wAlB-SG infected male Ae. aegypti at home address will be considered as a binary classification based on time since intervention (see intervention strategies) Outcome variable: test-positive status for dengue at point of testing Doubly robust logistic regression will be used to assess the intervention effect by estimating the aggregate odds ratio, which compares the exposure odds among test-positive cases vs test-negative controls. The null hypothesis is that the odds of residing in the intervention locations are the same among test-positive cases as test-negative controls All analyses will be adjusted for environmental and anthropogenic risk factors associated with dengue transmission risk, and the propensity to receive treatment based on the same risk factors. See detailed statistical analysis plan below | Same as hypothetical trial |
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
