We evaluated how a constrained supply of gonorrhoea POCTs could be allocated across testing strategies and priority populations in Kenya to maximise epidemiologic and health impact. Introducing 25,000 POCTs annually (sufficient to test 0.1% of adults) achieved the largest overall morbidity reduction when used for diagnostic confirmation testing among symptomatic pregnant women (3.5% of QALY losses averted). Diagnostic confirmation averted the most transmission when conducted among symptomatic FSW (2.1% of incident infections) or their male clients (2.2%), though with smaller morbidity reductions than focusing on pregnant women. Screening routine health service attendees was less effective (<0.6% of infections or QALY losses averted) than diagnostic confirmation for all priority populations. With unrestricted POCTs, screening yielded larger absolute effects, particularly among total population men and CFSW, but had lower per-test returns than diagnostic confirmation.

A constrained POCT supply should be prioritised for diagnostic confirmation over screening, supporting WHO guidance to strengthen aetiologic diagnosis within syndromic management. Diagnostic confirmation was more efficient than screening because symptomatic care attendees had higher per-test positivity than routine service attendees. With unrestricted tests, screening averted more total infections and QALY losses by detecting asymptomatic infections that would otherwise remain untreated, but it required substantially more tests to achieve these gains and had lower per-test returns. The benefit of unrestricted screening was greatest among larger population groups, but is unlikely to be feasible during early implementation.

Programmes aiming to maximise morbidity reduction with constrained POCTs should prioritise diagnostic confirmation among symptomatic pregnant women. Testing this group yielded the most health gains per test used, despite relatively low infection incidence, as untreated infections in pregnancy caused disproportionate QALY losses. However, focusing exclusively on pregnant women limits opportunities for broader infection control. In settings with increasing gonorrhoea prevalence, strategies targeting incidence reduction may therefore take precedence, whereas in stable or declining epidemics, prioritising high-burden groups offers greater immediate health gains.

Programmes aiming to maximise incidence reduction with constrained POCTs should prioritise diagnostic confirmation among higher-risk populations with symptomatic infection. Testing FSW or their clients was more efficient than testing lower-risk groups, due to higher prevalence and greater indirect benefits from reduced onward transmission. Simultaneously targeting FSW and CFSW may however yield diminishing returns at high coverage due to overlapping sexual networks. As POCT availability increases and coverage among high-risk groups saturates, our results suggest that diagnostic confirmation should expand to lower-risk symptomatic males. We evaluated interventions delivered alternately to each population; future modelling should assess combined and sequential test allocation.

Diagnostic confirmation averted more infections among male than female subgroups, despite the already high sensitivity of syndromic management for urethral discharge (80–90%)³⁷ compared to POCTs meeting the WHO Optimal TPP (95%). This reflects higher symptom presentation and care-seeking among males, together with fewer female syndromic cases attributable to gonorrhoea. Multiplex POCTs may have higher diagnostic utility for females, since vaginal discharge often has multiple aetiologies. As new platforms approach market introduction, their performance and optimal targeting should be re-evaluated.

Screening asymptomatic populations offered limited benefit, especially in low-prevalence groups. Even highly sensitive and specific POCTs yield low positive predictive values in such contexts, leading to false positives and unnecessary treatment, which may collaterally increase risk of AMR. Although we did not model false-positive treatment or resistance dynamics, evidence suggests intensified screening can drive resistance through greater antibiotic use.³⁸ Conversely, diagnostic confirmation would reduce overtreatment,¹⁴ as syndromic management has poor specificity for both urethral (60-70%)³⁷ and vaginal discharge (55–60%).⁸ These factors support prioritising POCTs for diagnostic use, especially where overtreatment and antibiotic exposure are common. WHO TPPs for POCTs detecting gonorrhoea resistance markers could enable more targeted treatment.³¹ Future models should incorporate resistance dynamics to align POCT deployment with antimicrobial stewardship.

Data limitations affected both calibration and parameterisation of the transmission model. Population-representative gonorrhoea prevalence estimates were scarce, and data for FSW largely came from urban cohort studies. Gaps in prevalence data required calibration based on male-to-female ratios for men and CFSW, and prevented direct calibration for AGYW and pregnant women. Sexual behaviour and care-seeking parameters relied on self-reported data, which are subject to reporting biases.³⁹ Parameters for CFSW and intermediate risk groups relied on sparse data or assumptions.

Structural assumptions in the transmission model may have influenced intervention impact estimates. We assumed sexual mixing proportionate to numbers of sexual contacts, which may have under-represented the concentration of transmission in younger and higher-risk populations.⁴⁰ We also did not incorporate heterogeneity in the frequency and exclusivity of mixing between FSW and CFSW. These assumptions are unlikely to have affected prioritisation rankings of FSW, CFSW, or pregnant women given large and contrasting differentials between these groups in incidence and morbidity. We excluded post-infection immunity, for which evidence is inconsistent,^41–43 enabling rapid reinfection that may have underestimated the impact of POCTs on transmission.⁴³ We assumed diagnosis depended solely on diagnostic sensitivity and was followed by universal access to effective treatment, overlooking barriers such as stock-outs or informal antibiotic use,⁴⁴ which may have overestimated the incremental benefit of POCTs over syndromic management. Our modelled screening scenarios did not incorporate delivery constraints, such as outreach to identify CFSW or ANC screening occurring at a single visit rather than uniformly across gestation, which could affect the relative costs and feasibility of intervention delivery. We assumed stable baseline gonorrhoea prevalence and incidence pre-intervention, a simplification of longer-term trends but consistent with relatively stable estimates in Eastern Africa.⁵ Although biological and behavioural factors suggest AGYW may have higher prevalence than older women, limited and mixed evidence across SSA precluded age differentiation,^45–47 potentially underestimating the impact of testing AGYW. MSM were not included despite high gonorrhoea prevalence in Kenya,⁴⁸ though our findings for other high-risk populations suggest prioritising MSM would be efficient for reducing transmission but less for averting population health losses. Finally, behavioural responses to POCT intervention, such as partner notification, altered care-seeking, or stigma were not modelled, though could affect uptake and effectiveness.

Estimating gonorrhoea-related health losses required several assumptions. Sequelae probabilities were primarily derived from chlamydia studies, potentially misrepresenting gonorrhoea-specific risks.²⁴ Several pregnancy-related outcomes, such as premature rupture of membranes and preterm birth, were excluded due to complex interactions with modelled sequelae and insufficient evidence for parameterisation,^2,3 likely underestimating the total burden. For stillbirths, we assigned a utility loss equivalent to neonatal death, though quantification lacks consensus:²⁸ some studies exclude stillbirth, while others assign lower losses to reflect the gradual accrual of life potential in early childhood.^28,49 Our sensitivity analyses applying these alternatives did not however change the prioritisation of pregnant women for reducing morbidity.

In conclusion, gonorrhoea POCTs meeting the WHO Optimal TPP should be deployed for diagnostic confirmation rather than screening when supply is constrained in Kenya, supporting WHO guidance to strengthen aetiologic diagnosis within syndromic management.⁶ Programmes should prioritise symptomatic pregnant women to maximise morbidity reduction or high-incidence populations such as FSW and their clients to maximise transmission reduction, depending on programmatic objectives. POCTs were more efficient for diagnosing urethral than vaginal discharge, even though vaginal discharge syndrome has poorer performance for gonorrhoea diagnosis. As new POCTs approach market introduction, particularly those with multiplex or resistance detection capabilities, prioritisation strategies should be re-evaluated to reflect diagnostic performance, epidemic dynamics, and the potential for combined targeting approaches.