Draft — LMIC Conference Grants

MOTIVATION LETTER

The European Health Psychology Society's LMIC Conference Grants exist to bring health psychology research and practice to countries where the burden of disease is highest and the research infrastructure thinnest. Nigeria is on the ISI list of eligible LMIC countries, and Nigeria carries a double burden: the highest HIV/TB/malaria caseloads in sub-Saharan Africa and a rising antimicrobial resistance crisis that the WHO has classified as a top-ten global health threat. My venture addresses the AMR half of that burden with a computational tool that predicts drug resistance mutations from protein sequence alone, and I am applying for this grant to attend the EHPS conference to learn the implementation science and behavioral change methods that determine whether a diagnostic tool actually changes prescriber behavior in a low-resource clinic.

The venture's technology is already validated at the proof-of-concept stage. It uses ESM-2 protein language model delta-embeddings combined with ECFP4 drug fingerprints, fed into a Random Forest classifier. On the Platinum benchmark of 553 mutations with protein-grouped cross-validation, it achieves an AUROC of 0.804 plus or minus 0.025, beating the published state of the art, mCSM-lig, which sits near 0.70. It covers 100 percent of mutations in the benchmark, where structure-limited tools cover roughly 18 percent because they require a crystal structure that does not exist for most clinically relevant proteins. On SKEMPI 2.0 it scores 0.634, which is the current gap the roadmap targets: fine-tuning ESM-2 on the SKEMPI 3K mutation set to push that number to 0.70 or higher.

The grant's purpose is conference attendance, and I am not misrepresenting it as venture funding. What I need from EHPS is method, not capital. My training is pharmacy and machine learning engineering, not health psychology. The tool predicts resistance; it does not by itself change the behavior of a clinician in Lagos who has limited time, limited formulary options, and a patient who cannot afford a second-line drug. The EHPS conference is the venue where I can learn the implementation science frameworks, the behavior change technique taxonomies, and the intervention design methods that turn a predictive model into a deployed health intervention. The collective conference report for the European Health Psychologist and the individual 300 to 500 word report on career and networking impact are commitments I will meet in full.

The career development case is concrete. I am a pharmacist-turned-ML engineer, sole founder of a pre-seed venture that is not yet incorporated. I have no formal training in health psychology, and my network in that field is zero. The EHPS conference would connect me to researchers who study clinician decision-making, antibiotic prescribing behavior, and patient adherence in LMIC settings. Those are the exact people whose methods I need to adapt for the venture's pilot work. The venture has a named partnership pipeline that includes Servier in Suresnes, Paris-Saclay's I2BC, Institut Pasteur, and Sanofi in Gentilly, but none of those partners specialize in behavioral science. EHPS fills that gap.

The relevance to health psychology is direct. Antimicrobial resistance is a behavior problem, not primarily a molecular biology problem. Resistance emerges when clinicians overprescribe, when patients complete courses incompletely, when diagnostics are ignored because they are slow or expensive, and when health systems lack the feedback loops that connect prescribing data to resistance surveillance. My tool makes resistance prediction fast and structure-free, but adoption in a Nigerian clinic depends on whether clinicians trust it, whether it fits their workflow, and whether it changes their prescribing decision. Those are health psychology research questions, and they are the questions I want to bring to the EHPS community.

The scientific merit of the underlying work is documented. The AUROC of 0.804 on Platinum with protein-grouped cross-validation is a published benchmark result, not an internal claim. The 100 percent mutation coverage versus 18 percent for structure-limited tools is a structural advantage of the sequence-only approach. The roadmap is explicit: fine-tune ESM-2 on SKEMPI 3K, reach AUROC 0.70 or higher on that harder benchmark, then run a pilot with Servier. The EHPS grant does not fund that roadmap, but it funds the conference attendance that equips me to design the behavioral component of the pilot correctly the first time.

I am requesting support to attend the EHPS conference, to learn implementation science and behavior change methods, and to return to Nigeria with a plan for integrating those methods into an AI-driven AMR diagnostic. The grant criteria ask for relevance to health psychology, scientific merit, career development potential, and clarity of proposed participation. My application addresses all four with specific evidence. I am prepared to submit electronically to grants@ehps.net and to commit to the collective and individual reports upon award.

Eniola Olutogun

MOTIVATION LETTER

The European Health Psychology Society's LMIC Conference Grants exist to bring health psychology research and practice to countries where the burden of disease is highest and the research infrastructure thinnest. Nigeria is on the ISI list of eligible LMIC countries, and Nigeria carries a double burden: the highest HIV/TB/malaria caseloads in sub-Saharan Africa and a rising antimicrobial resistance crisis that the WHO has classified as a top-ten global health threat. My venture addresses the AMR half of that burden with a computational tool that predicts drug resistance mutations from protein sequence alone, and I am applying for this grant to attend the EHPS conference to learn the implementation science and behavioral change methods that determine whether a diagnostic tool actually changes prescriber behavior in a low-resource clinic.

The venture's technology is already validated at the proof-of-concept stage. It uses ESM-2 protein language model delta-embeddings combined with ECFP4 drug fingerprints, fed into a Random Forest classifier. On the Platinum benchmark of 553 mutations with protein-grouped cross-validation, it achieves an AUROC of 0.804 plus or minus 0.025, beating the published state of the art, mCSM-lig, which sits near 0.70. It covers 100 percent of mutations in the benchmark, where structure-limited tools cover roughly 18 percent because they require a crystal structure that does not exist for most clinically relevant proteins. On SKEMPI 2.0 it scores 0.634, which is the current gap the roadmap targets: fine-tuning ESM-2 on the SKEMPI 3K mutation set to push that number to 0.70 or higher.

The grant's purpose is conference attendance, and I am not misrepresenting it as venture funding. What I need from EHPS is method, not capital. My training is pharmacy and machine learning engineering, not health psychology. The tool predicts resistance; it does not by itself change the behavior of a clinician in Lagos who has limited time, limited formulary options, and a patient who cannot afford a second-line drug. The EHPS conference is the venue where I can learn the implementation science frameworks, the behavior change technique taxonomies, and the intervention design methods that turn a predictive model into a deployed health intervention. The collective conference report for the European Health Psychologist and the individual 300 to 500 word report on career and networking impact are commitments I will meet in full.

The career development case is concrete. I am a pharmacist-turned-ML engineer, sole founder of a pre-seed venture that is not yet incorporated. I have no formal training in health psychology, and my network in that field is zero. The EHPS conference would connect me to researchers who study clinician decision-making, antibiotic prescribing behavior, and patient adherence in LMIC settings. Those are the exact people whose methods I need to adapt for the venture's pilot work. The venture has a named partnership pipeline that includes Servier in Suresnes, Paris-Saclay's I2BC, Institut Pasteur, and Sanofi in Gentilly, but none of those partners specialize in behavioral science. EHPS fills that gap.

The relevance to health psychology is direct. Antimicrobial resistance is a behavior problem, not primarily a molecular biology problem. Resistance emerges when clinicians overprescribe, when patients complete courses incompletely, when diagnostics are ignored because they are slow or expensive, and when health systems lack the feedback loops that connect prescribing data to resistance surveillance. My tool makes resistance prediction fast and structure-free, but adoption in a Nigerian clinic depends on whether clinicians trust it, whether it fits their workflow, and whether it changes their prescribing decision. Those are health psychology research questions, and they are the questions I want to bring to the EHPS community.

The scientific merit of the underlying work is documented. The AUROC of 0.804 on Platinum with protein-grouped cross-validation is a published benchmark result, not an internal claim. The 100 percent mutation coverage versus 18 percent for structure-limited tools is a structural advantage of the sequence-only approach. The roadmap is explicit: fine-tune ESM-2 on SKEMPI 3K, reach AUROC 0.70 or higher on that harder benchmark, then run a pilot with Servier. The EHPS grant does not fund that roadmap, but it funds the conference attendance that equips me to design the behavioral component of the pilot correctly the first time.

I am requesting support to attend the EHPS conference, to learn implementation science and behavior change methods, and to return to Nigeria with a plan for integrating those methods into an AI-driven AMR diagnostic. The grant criteria ask for relevance to health psychology, scientific merit, career development potential, and clarity of proposed participation. My application addresses all four with specific evidence. I am prepared to submit electronically to grants@ehps.net and to commit to the collective and individual reports upon award.

Eniola Olutogun

AI Draft — LMIC Conference Grants