Submission ID 131282
| Issue/Objective | Implementation research in LMICs is predominantly disease-specific, with findings reported vertically to programme funders rather than shared laterally across health systems. This produces a fragmented evidence base that obscures the system-level determinants of implementation success. Pakistan presents an unusual opportunity: concurrent implementation research programmes across maternal health, infectious disease, and mental health, operating within the same health system infrastructure. This paper extracts cross-cutting implementation determinants from this portfolio, arguing that lateral knowledge transfer - not disease-specific optimisation - generates the highest implementation value. |
|---|---|
| Methodology/Approach | The analysis draws on four concurrent implementation programmes: safe antenatal corticosteroids (ACS) use for early preterm births across five high-risk countries (BMGF, 2022-ongoing); integrated TB-diabetes care using a multiphase mixed-methods design (NIHR POTENTIAL study, protocol published BMJ Open 2024); mental health integration into diabetes services (NIHR IMPACT project, process evaluation lead); and opportunistic cervical cancer screening using VIA and PCR across three Islamabad hospitals (2025). Cross-programme analysis applied an implementation science framework examining determinants at health worker, facility, and system levels. |
| Results | hree implementation determinants emerged consistently across all programmes, irrespective of disease domain. Health worker cognitive load when adopting complex or dual-condition protocols was the primary adoption barrier - more predictive than training exposure or clinical complexity. Facility-level supervisors functioned as implementation accelerators or blockers disproportionate to their formal authority. Co-designed interventions demonstrated significantly stronger protocol fidelity at scale than those designed externally and contextually adapted. Supply chain reliability and disease-specific factors, by contrast, showed variable predictive value across programmes. |
| Discussion/Conclusion | These findings challenge disease-siloed implementation research commissioning. A cross-programme synthesis surfaces system-level insights - on supervision architecture, co-design methodology, and cognitive demand - that no single study could generate. The paper proposes a learning health system architecture enabling implementation findings to circulate laterally across programmes, with governance and financing implications for how ministries of health and global funders commission and institutionalise implementation research. Pakistan's portfolio offers a transferable model for LMICs managing concurrent GHI and bilateral research investments. |
| Presenters and Affiliations | Mariyam Sarfraz Health Services Academy Mariyam Sarfraz Health Services Academy |
