ATAcare exists because chronic disease falls hardest on populations the health system was not built around...

This piece lays out what the evidence currently shows...

The pattern, before the diseases

Across healthcare AI broadly, there is a well-documented split...

Put simply: the places where a tool is tested are rarely the places where it most needs to work.

Cardiovascular disease: trained on the wrong population

Cardiovascular risk prediction is one of the most mature applications...

The same review examined a separate cohort of 109,490 patients...

None of this is happening because anyone set out to build an unfair tool...

Diabetes: when the device itself reads the body differently

Diabetes management technology shows the same pattern from a different angle...

A 2025 study using continuous glucose monitoring data...

The hardware problem compounds this...

Sickle cell disease: built on the same flawed sensors

Sickle cell disease is where these two problems converge most directly...

AI is genuinely advancing in this space...

But a 2024 systematic review covering studies on AI/ML applications...

This is not a future risk...

Why this is an equity problem, not just a technical one

It would be easy to read all of this as a software engineering issue...

The patients underrepresented in cardiovascular and diabetes AI training data...

A model that has not been validated on a population is not a neutral tool...

What ATAcare is building toward

This series exists because ATAcare’s research mission is not separate from its clinical mission...

Tracking the evidence as it develops...

Naming the gap before our communities encounter it...

Building toward community-side evaluation capacity...

Connecting research to the people it is about...

The next installment looks at the layer underneath the algorithm: the equipment that never arrives or arrives broken, the basic diagnostic infrastructure that is itself missing in many of the settings ATAcare works in, and the medication access and pharmacogenomic gaps that determine whether care happens at all, before any question of AI bias becomes relevant.

• Mitigating the risk of artificial intelligence bias in cardiovascular care. The Lancet Digital Health, 2024.
• Artificial intelligence bias in the prediction and detection of cardiovascular disease. npj Cardiovascular Health, 2024.
• Overcoming AI Bias: Understanding, Identifying and Mitigating Algorithmic Bias in Healthcare. Accuray, 2024.
• Racial disparities in continuous glucose monitoring-based 60-minute glucose predictions among people with type 1 diabetes. medRxiv / PMC, 2024–2025.
• Algorithmic Bias in AI-Based Diabetes Care: Systematic Review of Model Performance, Equity Reporting, and Physiological Label Bias. 2025.
• Obermeyer, Z., Powers, B., Vogeli, C., Mullainathan, S. Dissecting racial bias in an algorithm used to manage the health of populations. Science, 2019.
• Sjoding, M.W., Dickson, R.P., Iwashyna, T.J., Gay, S.E., Valley, T.S. Racial bias in pulse oximetry measurement. New England Journal of Medicine, 2020.
• Pulse Oximeter Accuracy and Racial Bias. The Cardiology Advisor, 2025; FDA draft guidance, January 2025.
• Enhanced AI-Driven Prediction of Vaso-Occlusive Crises in Sickle Cell Disease. Blood (ASH 2024 abstract), ScienceDirect.
• Exploring machine learning algorithms in sickle cell disease patient data: A systematic review. PLOS One, 2024.
• Diagnostic accuracy and challenges of AI/ML models for sickle cell disease detection in blood smears: A meta-analysis and systematic review. ScienceDirect, 2025.
• Identifying Elevated Risk for Future Pain Crises in Sickle-Cell Disease Using Photoplethysmogram Patterns Measured During Sleep. Frontiers in Digital Health, 2021.
• Bridging the Gap: From AI Success in Clinical Trials to Real-World Healthcare Implementation. Narrative review, 2025.
• Deploying medical AI in low-resource settings: a scoping review of challenges and strategies. Frontiers in Digital Health, 2026.

The first post in this series asked whether AI tools for cardiovascular disease, diabetes, and sickle cell disease work as well for the patients ATAcare serves as they do in published benchmarks. That question assumes something most benchmark studies never have to think about: that the device exists, that it works, that someone can use it, and that the medication it leads to is actually available — in the right dose, for the right biology.

In much of West Africa, and in under-resourced clinics everywhere, none of that can be assumed. This post is about the layer underneath the algorithm — the equipment that never arrives, the equipment that arrives and breaks, and the medication that exists but does not reach the patient, or reaches them at the wrong dose for reasons nobody adjusted for. This is not a hypothetical concern. It is what determines whether a diagnosis happens at all, before any question of AI bias ever becomes relevant.

Equipment that never arrives, or arrives and breaks

The scale of this problem is larger than most people realize. A widely cited assessment of medical equipment access in low- and middle-income countries concluded that most medical devices transplanted from the developed world to the developing world hospital fail, with the WHO estimating that 70% of medical devices designed for developed-world use do not function as intended when they reach lower-resource settings (Malkin, Expert Review of Medical Devices, 2007). A follow-up study that examined 112,040 pieces of equipment across hospitals in sixteen countries quantified this precisely: an average of 38.3% of medical equipment in developing countries was out of service at the time of inspection, with lack of training, health technology management, and infrastructure identified as the three main causes (Perry and Malkin, Medical and Biological Engineering and Computing, 2011). More recent systematic reviews confirm that healthcare facilities in low-resource settings in sub-Saharan Africa continue to be “plagued with issues of non-functional and obsolete medical devices,” and that medical device design approaches have broadly failed to account for the contexts in which devices are actually used, with over 80% of the global medical device market share concentrated in high-resource settings that set standards others cannot take for granted (Piaggio et al., Globalization and Health, 2021; Maccaro et al., Frontiers in Sustainability, 2023).

This is not an abstract statistic to me. Through my work with Rayo Contra Cancer, an Einstein College of Medicine-affiliated nonprofit that provides radiotherapy equipment and technical training to English-speaking East African countries, I have seen this pattern firsthand. Cancer centers receive donated radiotherapy machines, and those machines can sit unused for years — not because the technology failed, but because a single small part broke and there was no trained technician in the country to replace it, and no supply chain to get the part there even if there had been. A multi-million-dollar piece of equipment, rendered useless by a five-dollar component and the absence of one trained person. The same paper notes that brain drain — skilled workers emigrating from their home countries — compounds this, because modern medical equipment requires highly skilled technicians to operate and maintain, and hospitals can be reluctant to invest in training staff who may then leave.

The lesson generalizes directly to ATAcare’s focus areas. The same structural gap — equipment without maintenance infrastructure — applies just as much to a glucometer in a rural Burkinabè clinic as it does to a linear accelerator in a regional cancer center. The scale and stakes differ; the underlying failure mode does not.

The diagnostic floor is missing before AI even enters the picture

Post 1 discussed AI tools that have not been validated on African populations. A deeper problem sits underneath that one: in many of the settings ATAcare works in, the basic diagnostic infrastructure needed to generate the data an algorithm would even use is itself missing.

Point-of-care diagnostic availability for hypertension and diabetes remains highly inconsistent across sub-Saharan Africa, with cost, supply chain reliability, and the lack of trained staff to operate and interpret testing equipment cited consistently as limiting factors. A recent cardiovascular health review noted plainly that existing cardiovascular risk prediction scores have not been validated in Sub-Saharan African contexts — which is not just an algorithmic bias problem, but partly a consequence of the fact that the underlying clinical data infrastructure to do that validation has never existed at scale in the region.

Sickle cell disease shows the same pattern in sharper relief. West Africa carries one of the highest sickle cell disease burdens in the world, yet systematic newborn screening — standard practice in high-income countries — remains absent across most of the continent. A six-year point-of-care screening program at Koutiala Hospital in rural Mali, using the HemotypeSC™ rapid test, demonstrated that low-cost, equipment-light testing is technically achievable at scale: over 18,000 newborns were screened with 99.2% coverage, and 95 infants with sickle cell disease were enrolled in the hospital’s treatment program who might otherwise never have received a diagnosis (Guindo, Cablay et al., British Journal of Haematology, 2025). This is precisely the kind of ground-up diagnostic infrastructure that makes downstream care — and eventually AI-assisted monitoring — possible at all.

Medication: available in theory, wrong in practice

Even when a diagnosis is successfully made, the medication that follows is its own access and appropriateness problem — and this is where ATAcare’s work touches the question most personally.

Hydroxyurea has been listed as a WHO essential medicine since 2013. It is the primary disease-modifying treatment for sickle cell disease, and economic modeling of its use in a low-income African setting has found it to be cost-saving relative to no treatment — meaning it is not just clinically effective, it is the cheaper option over a patient’s lifetime. And yet it remains largely inaccessible to the children who need it most across sub-Saharan Africa, limited by cost, supply, and the fact that safe use requires regular blood monitoring that many clinics cannot provide. A qualitative study of manufacturers, importers, and regulators in Tanzania documented these barriers in detail, finding supply discontinuities and regulatory constraints that mean the drug is simply unavailable at many points of care even when a clinician would prescribe it (Mlyuka HJ, Kilonzi M, Mutagonda RF et al., Healthcare, 2022;10(11):2223).

That monitoring gap changes how the drug is even prescribed. Because routine titration to maximum tolerated dose requires lab capacity many African clinics do not have, common practice in several countries is to give a fixed dose instead — a workaround driven entirely by the equipment and monitoring access problem described above, not by clinical preference.

Layered on top of that is a question almost no one is asking yet: does hydroxyurea work the same way for every patient, regardless of where their ancestors are from? The evidence says no. Up to 30% of sickle cell patients do not respond adequately to hydroxyurea, and researchers increasingly believe genetic factors are part of the reason — variants in genes including BCL11A, KLF10, CYP2C9, and CYP2E1 involved in fetal hemoglobin production and drug metabolism. A 2025 study of 148 sickle cell patients in Tanzania investigated exactly this question, identifying significant associations between hydroxyurea response and multiple genetic loci — but as the study’s own authors note, most prior research into the genetics of hydroxyurea response was conducted in non-African populations. The drug most central to sickle cell treatment in Africa has been pharmacogenomically studied mostly everywhere except Africa (Nkya S, Nzunda C, Saukiwa E et al., The Pharmacogenomics Journal, 2025;25(3):11. doi:10.1038/s41397-025-00372-3).

This pattern is not unique to hydroxyurea. The CYP2D6 enzyme alone metabolizes roughly a quarter of all prescribed drugs, and the frequency of gene variants that change how fast a person processes a drug differs sharply by ancestry — ultra-rapid metabolizer variants appear in up to 29% of some Ethiopian populations studied, compared to 1–2% in Northern European populations studied. For tamoxifen, a common breast cancer drug, an estimated 14–34% of people of African descent carry a variant that reduces the drug’s effectiveness — and CYP2D6 genotyping is still not required before dispensing it in sub-Saharan Africa. Clinical dosing guidance for warfarin, one of the most widely used blood thinners in the world, is split by ancestry group specifically because the genetic algorithms developed and validated in non-African populations do not transfer reliably to patients of African ancestry.

Even one of the most prescribed drug classes for hypertension — ACE inhibitors and ARBs — shows a related pattern through a different mechanism. Clinical trials and observational data have consistently found a blunted blood pressure response in patients of African descent with ACE inhibitors compared to diuretics or calcium channel blockers, attributed to lower renin-angiotensin system activity and a tendency toward volume-driven hypertension. This finding has been reflected in hypertension guidelines for decades: as of the 2017 ACC/AHA guidelines and the 2020 International Society of Hypertension global guidelines, thiazide diuretics and calcium channel blockers were specifically recommended as first-line therapy for Black patients without comorbid heart failure or chronic kidney disease. It is worth noting that the 2025 ACC/AHA guideline update removed this race-specific language, and a 2026 systematic review of 24 worldwide guidelines found the underlying trial evidence for avoiding ACE inhibitors in Black patients to be inconsistent across studies — a reminder that even well-established guideline distinctions deserve continued scrutiny. What is not in dispute is the broader point: dosing and prescribing guidance for common drugs was built overwhelmingly on trial populations that looked nothing like the patients ATAcare serves.

These are not edge cases. They are common drugs, common conditions, and a research base that, time and again, did its work somewhere else.

For a family managing sickle cell disease, hypertension, or diabetes across Upper Manhattan, the South Bronx, and Ouagadougou, this means three separate things have to go right before a medication actually helps: it has to be available, it has to be affordable, and it has to be the right choice for that person’s biology — and right now, the research and supply infrastructure that would guarantee all three was largely built without this population in mind.

What ATAcare is building toward

A clinic-pharmacy model designed around the access gap, not around an assumption that it does not exist. ATAcare’s planned clinic-pharmacy presence in Ouagadougou is built on the premise that diagnosis, medication access, and adherence support need to live in the same place a patient can actually reach — rather than requiring a referral chain across facilities that may each have their own equipment and supply gaps.

At-home diagnostic kit distribution as a deliberate workaround to the equipment-access problem. Distributing blood pressure monitors and glucometers directly to patients, with education and follow-up built in, sidesteps the reality that the nearest clinic may not reliably have a working device of its own — and mirrors the same logic behind the low-cost, equipment-light sickle cell testing innovations described above.

Treating medication adherence support as inseparable from medication access. ATAcare’s adherence coaching work — checking in on whether a patient actually has and is taking their medication, not just whether it was prescribed — exists because availability and adherence are two different failure points, and a program that addresses only one of them misses half the problem.

Carrying the maintenance lesson from radiotherapy into chronic disease equipment. The single biggest lesson from working with Rayo Contra Cancer is that donating equipment without local technical capacity to maintain it is not a complete intervention — it is half of one. As ATAcare’s at-home monitoring program grows, that lesson shapes how we think about distribution: a blood pressure monitor that breaks in eight months with no way to repair or replace it locally is not meaningfully different from a monitor that was never sent at all.

Naming the pharmacogenomic gap rather than ignoring it. ATAcare cannot run pharmacogenomic trials on its own. But we can be honest, in our patient education and in our advocacy, that a standard medication dose was very likely developed and tested somewhere far from the patient taking it — and that this is a real, documented gap in the evidence, not a reason for non-adherence, but a reason clinicians and patients alike deserve better research investment in this area. Longer term, we see a role for ATAcare in building the clinical data foundation and academic partnerships that could one day support that research directly, starting with the patients and continuity of care we are already building toward.

The next installment in this series returns to the AI question directly: what it would actually take to validate a sickle cell crisis-prediction tool — the kind already showing promising results in pilot studies — on the population it is meant to serve, and what specific gaps, including some described in this post, stand in the way.

• Malkin RA. Barriers for medical devices for the developing world. Expert Review of Medical Devices. 2007;4(6):759–763. doi:10.1586/17434440.4.6.759
• Perry L, Malkin R. Effectiveness of medical equipment donations to improve health systems: how much medical equipment is broken in the developing world? Medical and Biological Engineering and Computing. 2011;49(7):719–722. doi:10.1007/s11517-011-0786-3
• Piaggio D, Castaldo R, Cinelli M, Cinelli S, Maccaro A, Pecchia L. A framework for designing medical devices resilient to low-resource settings. Globalization and Health. 2021;17:64. doi:10.1186/s12992-021-00718-z
• Maccaro A et al. The extent to which circular economy principles have been applied in the design of medical devices for low-resource settings in Sub-Saharan Africa: a systematic review. Frontiers in Sustainability. 2023. doi:10.3389/frsus.2023.1079685
• Obermeyer Z, Powers B, Vogeli C, Mullainathan S. Dissecting racial bias in an algorithm used to manage the health of populations. Science. 2019;366(6464):447–453. doi:10.1126/science.aax2342
• Sjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial bias in pulse oximetry measurement. N Engl J Med. 2020;383(25):2477–2478. doi:10.1056/NEJMc2029240
• Guindo A, Cablay K, Anderson AR et al. Systematic point-of-care newborn screening for sickle cell disease in rural Mali, West Africa. Br J Haematol. 2025;207(5):2118–2122. doi:10.1111/bjh.70158
• Mlyuka HJ, Kilonzi M, Mutagonda RF et al. Barriers and facilitators of availability of hydroxyurea for sickle cell disease in Tanzania: a qualitative study of pharmaceutical manufacturers, importers, and regulators. Healthcare. 2022;10(11):2223. doi:10.3390/healthcare10112223
• Nkya S, Nzunda C, Saukiwa E et al. Exploring pharmacogenetic factors influencing hydroxyurea response in Tanzanian sickle cell disease patients: a genomic medicine approach. The Pharmacogenomics Journal. 2025;25(3):11. doi:10.1038/s41397-025-00372-3
• Are Angiotensin-Converting Enzyme Inhibitors Effective in the Treatment of Hypertension in Black Patients? Systematic review of 24 worldwide guidelines. American Journal of Preventive Cardiology. 2026. doi:10.1016/j.ajpc.2026.100960
• Whelton PK et al. 2017 ACC/AHA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults. Journal of the American College of Cardiology. 2018;71(19):e127–e248.
• Unger T et al. 2020 International Society of Hypertension Global Hypertension Practice Guidelines. Hypertension. 2020;75(6):1334–1357.
• Rayo Contra Cancer. Organizational mission and program description. rayoscontracancer.org.

The first two posts in this series made a general argument: AI tools in chronic disease care are often trained, tested, and validated on populations that don’t reflect the patients who carry the highest burden of disease, and the basic equipment and medication infrastructure needed to generate reliable health data is itself missing or inconsistent across much of the world these patients live in. This post takes that argument and applies it to one specific, real, and genuinely promising technology: AI models that predict a sickle cell vaso-occlusive crisis before it happens.

This is not a hypothetical exercise. These models exist today, in pilot form, and the early results are encouraging. The question this post asks is narrow and concrete: what would it actually take to know whether one of these tools works for a patient in Ouagadougou, or the South Bronx, the way it appears to work in the cohort it was built on? And what, specifically, is currently missing to get there?

What these tools can already do

A vaso-occlusive crisis (VOC) is the acute pain episode that drives most sickle cell hospitalizations, and it remains a leading cause of organ damage and death in people living with the disease. Predicting one before it escalates to the emergency room has obvious value, and several research groups are actively building toward it.

A model presented at the American Society of Hematology’s 2024 annual meeting used an ensemble of machine learning methods — gradient boosting, neural networks, and clustering — to generate a daily 0–100% crisis-risk score from patient-reported symptoms and physiological data. An earlier version of the model was tested on 186 patients; an updated version expanded to 399 patients over six months. The variables that mattered most to the model were intuitive and clinically sensible: fatigue, pain score, and hydration-related factors. A separate line of research has gone further upstream, using overnight wearable sensors to detect autonomic nervous system changes — specifically vasoconstriction patterns measured through photoplethysmography (PPG) — that appeared predictive of a crisis days before it occurred.

These are real, meaningful steps. They also remain, by their own authors’ description, feasibility and pilot studies — proof that the approach is worth pursuing, not yet proof that it works reliably across the population it would need to serve.

The validation gap, in plain numbers

Here is the central problem: the published descriptions of these pilot cohorts do not report the racial, ethnic, or geographic composition of the patients involved. That silence matters, because the wearable hardware these models often depend on has a documented, independently measured accuracy gap by skin tone.

Photoplethysmography — the optical sensing method used in most consumer and clinical wearables to track heart rate, oxygen saturation, and the vasoconstriction signals being explored for VOC prediction — has been shown to produce inaccurate heart rate readings up to 15% more often in dark skin than in light skin, because melanin absorbs more of the light the sensor depends on. A more recent evaluation of consumer smartwatches found some devices underestimated heart rate by 10–15 beats per minute at rest in darker-skinned users, with errors exceeding 20% during physical activity. This is the same underlying sensor limitation discussed in Post 1 with respect to pulse oximetry — and sickle cell disease overwhelmingly affects populations of African and Mediterranean descent, the same populations these sensors are documented to read least reliably.

The scale of the testing itself compounds the concern. A scoping review of wearable PPG accuracy studies found that most published evaluations include fewer than 100 participants, and the review’s authors stated plainly that this sample size is too small to assess how a device performs across real-world diversity — including skin tone. One of the feasibility studies specifically exploring wearable PPG for VOC prediction enrolled 12 sickle cell patients and 6 healthy controls. That is an appropriate size for a first feasibility study. It is not a sample that can answer whether the underlying sensor technology performs consistently across the population sickle cell disease actually affects.

Put together: the device most of these prediction models depend on has a known, quantified, skin-tone-related accuracy problem, and the studies built on top of that device have not yet been run at a scale or with the reporting transparency needed to know whether that problem is showing up in their predictions.

A counterexample worth learning from

It is worth being fair here, because not every sickle cell prediction effort has this gap, and the contrast is instructive.

The PRESEV score, a clinical prediction tool for a related complication called acute chest syndrome, was developed from an initial cohort of 250 patients in France and explicitly designed for international validation. The follow-up validation study was conducted prospectively across thirteen centers on two continents, in five countries including Mali and Togo alongside France, Belgium, and England, enrolling 393 patients. This is precisely the kind of validation the AI-driven VOC models have not yet undergone: multi-country, multi-site, prospectively designed to test whether a tool developed in one population holds up in another (Kassasseya C et al., NEJM Evidence, 2026).

The difference is instructive. PRESEV2 used a small number of standard laboratory values — reticulocyte and white cell counts, hemoglobin, a categorical pain score — measurable with basic, widely available lab equipment, which made multi-country validation logistically achievable. The newer AI-driven VOC models depend on continuous wearable sensor data and daily patient-reported input, which raises the bar for what a validation study would need: not just access to diverse patients, but access to functioning wearable devices, reliable data transmission, and sustained patient engagement over months — the exact kind of infrastructure Post 2 described as inconsistent or absent across much of West Africa.

A tool’s predictive accuracy and a tool’s validation feasibility are two different problems, and the second one is often the harder one to solve in a low-resource setting.

What it would actually take

Given all of this, validating a sickle cell crisis-prediction tool on a population like ATAcare’s would require several concrete things, none of which currently exist together in one place:

Sensor performance testing stratified by skin tone, before clinical validation begins. Before any crisis-prediction model is tested on patients in Ouagadougou or the South Bronx, the underlying PPG hardware needs its own accuracy testing across the actual range of skin tones in those populations — using the same kind of polarization-sensitive sensor evaluation methods now emerging in the research literature, rather than assuming consumer-grade hardware performs equivalently for everyone.

A validation cohort large enough, and reported transparently enough, to answer the equity question. Twelve to 400 patients in a single-region pilot is appropriate for proving a concept. Knowing whether the tool works equitably requires a multi-site cohort, in the PRESEV2 tradition, with race, ethnicity, and geographic site reported as a matter of course — not omitted, as in the current VOC model publications.

Infrastructure for sustained data collection in lower-resource settings. A model that depends on nightly wearable wear, charged batteries, and continuous data transmission assumes a level of consistent connectivity and device maintenance that mirrors the same equipment-access problem described in Post 2. Any validation effort in West Africa would need to either adapt the model to lower-infrastructure data inputs, or pair it with the kind of device distribution and maintenance support ATAcare is already building toward.

A local clinical and research partner who can sustain the relationship. Multi-country studies like PRESEV2 worked because they had committed clinical sites in each country, not just patients enrolled remotely. ATAcare’s clinic-pharmacy model, and the relationships already being built with hematologists and nurses in Ouagadougou, are the kind of standing local infrastructure that a future validation study would actually need — not as an afterthought, but as the foundation that makes the validation possible at all.

What ATAcare is building toward

This series has tried to avoid promising more than is actually being built. So, plainly: ATAcare is not currently running a sickle cell crisis-prediction validation study. What this post lays out is the specific, concrete shape of the work that would need to happen for one of these promising tools to be trustworthy for the patients ATAcare exists to serve — and the parts of that work ATAcare is positioned to contribute to as the organization grows.

The clinical relationships already forming in Ouagadougou — with hematologists, nurses, and families living with sickle cell disease day to day — are the same kind of standing local infrastructure that made PRESEV2 possible across five countries. The at-home monitoring program described in Post 2 is a direct, practical step toward the device-access and maintenance support any future validation effort would require. And the patient education and advocacy work ATAcare is already doing means that if and when one of these tools is ready for testing in West Africa, the patients and clinicians involved will understand exactly what is and is not yet known about whether it works for them.

The next installment in this series turns to a technology that is not hypothetical at all: gene therapy for sickle cell disease is already FDA-approved and being called a cure. This post has been about validating a promising tool before it reaches patients; the next one is about a treatment that already exists, and the reasons it remains effectively out of reach for the patients who need it most.

• Summers KZ, Agrippa O, Ade-Odunlade D, Anie KA, Telfer P, Lugthart S. Enhanced Artificial Intelligence (AI)-Driven Prediction of Vaso-Occlusive Crises in Sickle Cell Disease: Precision through Advanced Machine-Learning Frameworks and Digital Remote Monitoring. Blood. 2024;144(Supplement 1):522. doi:10.1182/blood-2024-209961
• Vuong C, Utkarsh K, Stojancic R, Subramaniam A, Fernandez O, Banerjee T, Abrams DM, Fijnvandraat K, Shah N. Use of consumer wearables to monitor and predict pain in patients with sickle cell disease. Frontiers in Digital Health. 2023;5:1285207. doi:10.3389/fdgth.2023.1285207
• Shah P, Sy S, Chen M, Khoo MCK, Coates TD, Veluswamy S. Wearable Device Photoplethysmography As a Viable Tool to Longitudinally Monitor Vasoconstriction Biomarkers for Predicting Vaso-Occlusive Crisis in Sickle Cell Disease: Feasibility and Validation Study. JMIR Human Factors. 2026;13:e75465. doi:10.2196/75465
• Bent B, Goldstein BA, Kibbe WA, Dunn JP. Investigating sources of inaccuracy in wearable optical heart rate sensors. npj Digital Medicine. 2020;3:18. doi:10.1038/s41746-020-0226-6
• Asif S, AlSaafeen A, Nadar S, Nambiar S, Dannawi J, Korrapati NH, Wilkhoo HS. Photoplethysmography in Diverse Skin Tones: Evaluating Bias in Smartwatch Health Monitoring. Cureus. 2025;17(10):e94074. doi:10.7759/cureus.94074
• Jakachira R et al. Evaluation of a Polarization-Sensitive, Dual-Wavelength Wearable Photoplethysmography Sensor Across a Range of Skin Tones. Biophotonics Discovery. 2025;3(1):012509. doi:10.1117/1.BIOS.3.1.012509
• Knight S, Lipoth J, Namvari M, Gu C, Hedayati CM, Syed-Abdul S, Spiteri RJ. The Accuracy of Wearable Photoplethysmography Sensors for Telehealth Monitoring: A Scoping Review. Telemedicine and e-Health. 2022;29:813–828. doi:10.1089/tmj.2022.0182
• Kassasseya C, Kéné S, Besse-Hammer T, Nzouakou R, Magnang H, Telfer P, Arlet JB et al. Validation and Application of a Predictive Score of Acute Chest Syndrome. NEJM Evidence. 2026;5(1):EVIDoa2500074. doi:10.1056/EVIDoa2500074
• NCT03032055. Validation of a Predictive Score of Acute Chest Syndrome (PRESEV2). ClinicalTrials.gov.
• Madhi F, Kamdem A, Jung C, Carlier-Gonod A, Biscardi S, Busca J, Arnaud C, Hau I, Narbey D, Epaud R et al. Identification of Clinical and Laboratory Parameters Associated with the Development of Acute Chest Syndrome during Vaso-Occlusive Episodes in Children with Sickle Cell Disease. Journal of Clinical Medicine. 2019;8(11):1839. doi:10.3390/jcm8111839

In December 2023, the United States Food and Drug Administration approved two cell-based gene therapies for sickle cell disease — Casgevy and Lyfgenia — marking the first time CRISPR gene-editing technology received regulatory approval for any disease. For families who have watched sickle cell disease shape their children’s lives, hospitalization by hospitalization, crisis by crisis, the word attached to these approvals was “cure.” That word is not wrong, but it requires context that most headlines did not provide.

This post is about what the approval of these therapies actually means, who it currently reaches, why the gap between regulatory approval and patient access is itself a form of inequity, and where ATAcare is actively positioning itself to help close that gap — starting with what is possible today, and building deliberately toward what must become possible tomorrow.

What these therapies actually do

Both Casgevy and Lyfgenia are one-time treatments that work by modifying a patient’s own blood stem cells. In Casgevy, CRISPR/Cas9 gene-editing technology is used to reactivate fetal hemoglobin production — the fetal form of hemoglobin that, when present in sufficient quantities, prevents red blood cells from sickling. In Lyfgenia, a different approach is used: a modified gene is inserted that causes the patient’s blood cells to produce a functional hemoglobin variant called HbAT87Q, which performs similarly to normal adult hemoglobin. Both are approved for patients 12 years of age and older with a history of vaso-occlusive crises at the time of original approval (FDA, December 8, 2023) — expanded on July 1, 2026 to patients aged 2 years and older (FDA, July 1, 2026).

The clinical trial results are genuinely strong. In trials supporting approval, 93.5% of patients receiving Casgevy experienced no vaso-occlusive events in the months following treatment, and 88% of those receiving Lyfgenia were free from severe vaso-occlusion events. Because the treatment uses a patient’s own cells, there is no need for a donor match — one of the major barriers that has limited bone marrow transplantation, the previous best curative option, to a small minority of patients (Sickle Cell Disease Association of America, 2023). The FDA recommends 15 years of follow-up monitoring to assess long-term outcomes, meaning the word “cure” is used cautiously even by the SCDAA, which describes these therapies as “potentially curative.”

Who is actually receiving these therapies

Casgevy carries a list price of $2.2 million. Lyfgenia carries a list price of $3.1 million. Both require the patient’s stem cells to be harvested, sent to a manufacturer for modification, returned, and then infused following a course of high-dose chemotherapy to clear the bone marrow — a process that requires extended hospitalization, specialized care, and months of recovery. As of late 2024, when the first patients outside clinical trials began receiving treatment, there were approximately 50 authorized treatment centers for Lyfgenia and over 35 for Casgevy across the entire United States (Congressional Budget Office, December 2024).

Approximately 100,000 Americans live with sickle cell disease, and between 50% and 60% of them are insured through Medicaid — the public insurance program for low-income individuals. The Congressional Budget Office, analyzing coverage and access in detail, found that the five largest commercial insurers have established coverage criteria for both therapies that are more restrictive than the FDA’s indications, including prior authorization requirements, age restrictions, and requirements for prior hydroxyurea failure. A federal response has begun: the Centers for Medicare and Medicaid Services launched a voluntary Cell and Gene Therapy Access Model in January 2025, with 33 states plus Washington D.C. and Puerto Rico participating, in which the federal government negotiated outcomes-based pricing agreements with both manufacturers on behalf of state Medicaid programs. This is an important structural step. It does not yet mean the treatment is straightforwardly accessible. As one clinical director noted, even after successful negotiations, the real-world cost of a single course of treatment is likely to remain between $1 million and $1.5 million, and the infrastructure requirements — authorized treatment centers, stem cell collection, chemotherapy conditioning units, months-long recovery support — create barriers that pricing alone cannot solve.

The global access gap

Outside the United States and a small number of other high-income countries, the picture is fundamentally different. No African country — not one — currently has approved Casgevy or Lyfgenia through its national regulatory body, has an authorized treatment center, or has a financing mechanism that could make either therapy accessible at scale. West Africa, which carries one of the highest sickle cell disease burdens in the world, is outside this picture entirely.

A 2024 peer-reviewed analysis published in Frontiers in Genetics examined this gap explicitly, identifying the specific barriers: high treatment cost, the complexity of obtaining informed consent for minor patients, inadequate public health infrastructure, and the absence of regulatory oversight capacity in the countries where the disease burden is greatest. The authors frame these not as temporary administrative delays but as structural inequalities embedded in how global health governance works — where access depends on social and institutional arrangements that did not include the most affected populations when these therapies were developed (Lee and Sawai, Frontiers in Genetics, 2024).

A separate analysis published in Gene Therapy in 2026, modeling gene therapy cost-effectiveness specifically for Uganda, concluded that while these therapies could in theory be cost-effective even in a low-income African setting over a patient’s lifetime — given the enormous cost of managing severe, untreated sickle cell disease — the upfront price and the infrastructure required to deliver them make access effectively impossible without major structural intervention. The study’s own framing captures the situation precisely: “Gene therapy could have blockbuster drug potential if distributable to these regions, but cost is a major barrier.” The disease burden is in Africa. The treatment infrastructure is not.

The Gates Foundation has invested in developing a simpler, single-shot in vivo gene therapy that could eventually lower both cost and delivery complexity — an approach that, if it reaches clinical scale, could fundamentally change what is deliverable in lower-resource settings. These are investments worth watching closely, and ATAcare intends to watch them. The timeline to a child in Ouagadougou is long. That is not a reason to stop building toward it.

Where ATAcare positions itself — and where it is building toward

ATAcare’s mission is explicit: to improve health outcomes for underserved populations wherever they are. That means being honest about what exists today while actively building the pathways that bring new treatments closer to the patients who need them most. Gene therapy is not yet accessible to most of the families ATAcare serves. That is the honest starting point. It is not the destination.

Providing accurate patient and family education — now. Many families in West Africa and in diaspora communities in the South Bronx and Upper Manhattan hear “cure” from news coverage without understanding what the treatment pathway actually requires — the age threshold, the chemotherapy, the authorized treatment center, the months of recovery. Honest, accessible information in French and English about who is currently eligible, what the treatment involves, and what the realistic timeline for broader access looks like is something ATAcare can offer now, and it matters. A family that understands the landscape can advocate for itself; a family that does not cannot.

Navigating access for patients who already have a pathway. Not every patient ATAcare serves is in Burkina Faso. Many are in the South Bronx and Upper Manhattan, where Medicaid is the primary insurer and where the CMS Cell and Gene Therapy Access Model creates a real, if still difficult, pathway. Helping families in these communities understand whether they qualify, which treatment centers exist in the New York area, and how to initiate a referral conversation with their hematologist is a concrete service ATAcare can provide today — and will.

Tracking trials, compassionate-use programs, and expanding-access arms. Clinical trials for next-generation gene therapies, including in vivo approaches designed for lower-resource delivery, are ongoing. ATAcare will monitor these systematically and flag opportunities — including compassionate-use programs, patient assistance pathways, and international expanded-access efforts — for every family we serve, wherever they are.

Building toward the partnerships that close the global access gap. The structural barriers described in this post — cost, infrastructure, regulatory capacity — cannot be solved by any single organization alone. But they can be addressed through deliberate, sustained partnership: with academic medical centers that run international sickle cell trials, with global health organizations working on next-generation delivery models, with manufacturers developing outcomes-based access programs for low-income settings, and with the hematologists and community health workers already on the ground in Ouagadougou. ATAcare’s clinic-pharmacy presence, its bilingual patient education capacity, its relationships with families and clinicians in both New York and Burkina Faso, and its standing as a 501(c)(3) research and community health organization are the foundation on which those partnerships will be built. We are not yet the organization that delivers gene therapy to West Africa. We are building toward being the organization that, when the delivery becomes possible, is already trusted, already present, and already connected to the families who need it.

Grounding everything in what the work requires right now. For the overwhelming majority of patients ATAcare serves today — including children with sickle cell disease in Ouagadougou — the immediate interventions are hydroxyurea access, newborn screening, pain crisis management, medication adherence, and the kind of consistent primary care that prevents crises from escalating. ATAcare’s programs are built around those needs first. Gene therapy is a horizon we are actively moving toward — not passively watching — and everything we build today is designed to make that movement possible.

The next installment in this series turns from what medicine cannot yet deliver everywhere to what the health system fails to deliver even when it can: what happens when the same blood tests are ordered twice at the same clinic, a month apart, because no reliable record of the first visit exists.

• U.S. Food and Drug Administration. FDA Approves First Gene Therapies to Treat Patients with Sickle Cell Disease. Press release, December 8, 2023. fda.gov/news-events/press-announcements/fda-approves-first-gene-therapies-treat-patients-sickle-cell-disease
• U.S. Food and Drug Administration. FDA Approves First Gene Therapy for Young Children with Sickle Cell Disease. Press release, July 1, 2026. fda.gov/news-events/press-announcements/fda-approves-first-gene-therapy-young-children-sickle-cell-disease
• Vertex Pharmaceuticals. Vertex Announces US FDA Approval for Expanded Use of CASGEVY® for the Treatment of People Ages 2 Years and Older With Sickle Cell Disease or Transfusion-Dependent Beta Thalassemia. Press release, July 1, 2026. news.vrtx.com
• Sickle Cell Disease Association of America. SCDAA Statement About Gene Therapy Approval. sicklecelldisease.org, December 2023.
• Congressional Budget Office. How Increased Use of Gene Therapy Treatment for Sickle Cell Disease Could Affect the Federal Budget. CBO Publication 61149, December 2024. cbo.gov/publication/61149
• Centers for Medicare and Medicaid Services. Biden-Harris Administration Takes Next Steps to Increase Access to Sickle Cell Disease Treatments. CMS Press Release, December 2024. cms.gov
• Hasibuzzaman MA et al. FDA approval of Casgevy and Lyfgenia: a dual breakthrough in gene therapies for sickle cell disease. Annals of Medicine and Surgery. 2024;86(9):4966–4968. doi:10.1097/MS9.0000000000002409
• Lee TL, Sawai T. Navigating equity in global access to genome therapy: expanding access to potentially transformative therapies and benefiting those in need requires global policy changes. Frontiers in Genetics. 2024;15:1381172. doi:10.3389/fgene.2024.1381172
• Obeagu EI, John A. Health equity in sickle cell disease: overcoming barriers to care in marginalized communities. Annals of Medicine and Surgery. 2025. doi:10.1097/MS9.0000000000004195
• Gene therapy cost-effectiveness for sickle cell disease in Uganda: tailoring high-income evidence to Uganda’s context. Gene Therapy. 2026. doi:10.1038/s41434-026-00598-1

A few weeks ago, I brought my father to a clinic near our home in Ouagadougou. His chief complaint was fever and hemorrhoid pain that had persisted for several days. The clinic ordered a panel of tests: PSA, A1C, fasting glucose, and several others. These were the same tests, at the same facility, that had been run one month earlier at the same clinic during a previous visit.

The tests were not ordered because the results had changed or because his condition had evolved in a way that required re-evaluation. They were ordered because the clinic had no reliable way to access what had been done the month before. The record of the first visit had not followed my father into the second one, even within the same institution.

This is not a story about one clinic making one mistake. It is a description of how care fragmentation — the inability of health information to travel with a patient across visits, providers, and time — operates as a daily, invisible tax on patients with chronic disease everywhere the digital health infrastructure has not yet been built. In Ouagadougou, and across much of sub-Saharan Africa, that infrastructure is still being assembled.

What care fragmentation costs patients

For a patient managing hypertension, diabetes, or sickle cell disease, continuity of information is not an administrative convenience. It is a clinical requirement. A physician who does not know what was prescribed last month may prescribe something that interacts with it. A lab that does not know what was tested last month orders it again, which costs money the patient may not have, delays the appointment, and contributes nothing to understanding whether the patient’s condition is improving or worsening over time. A nurse who does not know what a patient was counseled about at a previous visit starts from the beginning rather than from where the patient actually is.

Multiplied across the millions of Africans managing chronic conditions — the WHO estimates that non-communicable diseases now account for over 37% of all deaths in sub-Saharan Africa, with cardiovascular disease and diabetes rising fastest — the cost of fragmented care is not just individual inconvenience. It is a structural inefficiency that undermines every other investment made in treatment, medication, and monitoring. You can stock a clinic with the right drugs and working equipment; if the clinician seeing the patient today does not know what the clinician last month prescribed, some of that investment is wasted.

The digital health landscape in sub-Saharan Africa: real progress, real limits

The state of digital health in sub-Saharan Africa is neither as bleak as it was a decade ago nor as coherent as it needs to be. A 2025 scoping review covering ten years of electronic health record implementation across the region — 30 studies from 2014 to 2024 — found steady progress in the digitization of patient records and reduction of manual documentation errors, alongside persistent barriers: unreliable electricity, insufficient internet connectivity, limited staff training, and, most importantly, interoperability problems that mean facilities using different systems often cannot exchange patient data even when both have electronic records (Mugauri HD et al., Global Health: Science and Practice, 2025).

OpenMRS, the leading open-source electronic health record platform in sub-Saharan Africa, has been deployed in over 80 countries. DHIS2, the health data management platform, is used for aggregate reporting in more than 40 African countries. These are real achievements. What they have not yet produced is a system in which a patient’s full clinical history follows them reliably from one facility to another, or from one visit to the next within the same facility, across the region.

The challenge is not ignorance of the problem. It is the accumulated complexity of building interoperability on top of fragmented, historically siloed digital infrastructure — hundreds of disparate applications funded by different international partners, each designed for a specific program or disease area, not for a unified patient record.

Burkina Faso specifically: a system in active transition

The situation in Burkina Faso is a particularly clear example of both how far the transition has come and how much remains unfinished. A 2024 assessment published in BMC Medical Informatics and Decision Making found that Burkina Faso’s digital health landscape was characterized by fragmented implementation, intervention duplication, and unequal distribution across regions. By 2020, more than 110 distinct digital applications had been identified operating within the national health system — most of them siloed, many of them funded by separate international partners with separate reporting requirements, and none of them connected to a unified patient record infrastructure (Dembélé AK et al., BMC Medical Informatics and Decision Making, 2024;24:171).

The national response began in 2013 with the establishment of ENDOS-BF, a national health data warehouse built on DHIS2. But ENDOS-BF captures aggregate data — population-level statistics, program outcomes, epidemiological surveillance — not individual patient records. A patient’s PSA result from a visit last month does not live in ENDOS-BF.

A newer effort has been underway since 2023: the Minimal Digital Ecosystem (MDE), introduced as part of a national strategy to rationalize the fragmented landscape. The MDE consists of eight core digital tools implemented synchronously across primary health facilities, community health workers, and district referral facilities. A 2026 study evaluating adoption of MDE tools among primary healthcare professionals in Burkina Faso found that implementation had reached comparable levels across the study districts by mid-2024 — a meaningful logistical achievement in a country managing significant security and infrastructure challenges (Diallo AH et al., BMC Health Services Research, 2026). What the MDE does not yet provide, and what the study’s own framework acknowledges, is seamless patient-level record sharing across facility types and between the primary and referral levels.

This is the system my father moves through when he seeks care in Ouagadougou. It is a system actively being built, not a system that has been abandoned. The work is real. So is the gap.

Why this matters for chronic disease specifically

Most of what the digital health literature discusses in the context of sub-Saharan Africa concerns HIV/AIDS, maternal health, malaria, and infectious disease surveillance — the disease areas that have received the most sustained international investment in health information infrastructure. Cardiovascular disease, diabetes, and sickle cell disease receive far less attention in this literature, and that asymmetry is itself significant.

Chronic disease management requires something HIV and maternal health programs also need but that is especially acute for hypertension, diabetes, and sickle cell care: longitudinal tracking. A blood pressure reading is not meaningful in isolation. An A1C is more informative when you can see the one from three months ago alongside it. A sickle cell patient’s crisis history, transfusion record, medication response, and laboratory trends need to travel with the patient — to the hematologist, the emergency room, the community pharmacy, the nurse doing an adherence check — or each provider starts from zero and the chronic management of the disease is replaced by episodic crisis response.

The posts in this series so far have described AI tools that don’t work as well for African patients, devices that misread dark skin, medications calibrated to different genetics, and equipment that breaks without anyone to fix it. Care fragmentation is the connective tissue underneath all of those problems: even when a diagnosis is made correctly, with working equipment, leading to the right medication at the right dose, if the record of that encounter doesn’t travel with the patient, the clinical benefit is at risk of being lost at the next visit.

What ATAcare is building toward

ATAcare cannot build Burkina Faso’s national health information infrastructure. That is the work of governments, international agencies, and sustained public investment, and it is already underway, however unevenly. What ATAcare can do — and what the clinic-pharmacy model is specifically designed to do — is address the care continuity problem for its own patients, within its own scope, while that larger infrastructure is being built.

Maintaining longitudinal patient records from the first contact. ATAcare’s clinic-pharmacy will keep structured, longitudinal records for every patient it serves — not just visit notes, but medication history, laboratory results, adherence data, device readings from at-home monitoring, and counseling records — so that each clinical interaction builds on the last one, regardless of which provider the patient sees within ATAcare’s own network.

Building records that patients can carry. Where digital interoperability doesn’t exist, patient-held records — a structured summary of current medications, recent laboratory values, and key clinical events that a patient carries to any provider — are a proven intermediate solution. ATAcare will develop bilingual (French and English) patient health summaries for its chronic disease patients in Ouagadougou, designed to be readable and useful to any provider the patient visits, inside or outside ATAcare’s network.

Contributing to rather than duplicating existing national efforts. ATAcare’s digital health approach will align with the MDE and ENDOS-BF frameworks where possible rather than adding to Burkina Faso’s proliferation of siloed applications. The goal is to fit into the national infrastructure being built, not to create a parallel one.

Naming care fragmentation explicitly in patient education. Patients and families managing chronic disease deserve to understand why they are sometimes asked to repeat tests they have already had, and what they can do about it — including keeping their own records, asking for written summaries of visits, and advocating for their own continuity of care in settings where the infrastructure does not yet do it for them.

The next installment in this series turns to a different kind of knowledge gap: the traditional and herbal medicine that tens of millions of West African patients use alongside, instead of, or before conventional treatment — and what a pharmacist-led organization like ATAcare is positioned to do with that reality.

• Mugauri HD, Chimsimbe M, Shambira G et al. A decade of designing and implementing electronic health records in Sub-Saharan Africa: a scoping review. Global Health: Science and Practice. 2025. doi:10.1080/16549716.2025.2492913
• Dembélé AK, Karambiri M, Baguiya A et al. Current status of digital health interventions in the health system in Burkina Faso. BMC Medical Informatics and Decision Making. 2024;24:171. doi:10.1186/s12911-024-02574-4
• Diallo AH et al. Factors associated with digital tools use among primary healthcare professionals in Burkina Faso: a cross-sectional study of the minimal digital ecosystem. BMC Health Services Research. 2026. doi:10.1186/s12913-026-14331-6
• World Health Organization. Noncommunicable diseases in the WHO African Region: situation and perspectives. WHO Regional Office for Africa, 2022.

Before hydroxyurea, before blood transfusions, before hematologists — people in West Africa were already treating sickle cell disease. They used Cajanus cajan (pigeon pea), Zanthoxylum zanthoxyloides (fagara), Carica papaya, and dozens of other plants documented across the region as antisickling remedies. They used bitter leaf (Vernonia amygdalina) and shea butter preparations for cardiovascular symptoms. They used traditional healers whose knowledge of local pharmacopoeia was precise, contextual, and accumulated over generations.

This is not an alternative history. It is the current reality for a very large number of patients. A 2024 systematic review of herbal medicine use in sickle cell disease across Africa found that traditional remedies remain widely used across the continent, with evidence that in many communities they are the first intervention tried before any conventional care, and in some communities the only one regularly accessible. Surveys across multiple countries document active, concurrent use of traditional and biomedical medicine — not as a failure of patient knowledge, but as a rational response to the constraints of cost, geography, and supply.

A pharmacist-led organization operating in Ouagadougou and the South Bronx cannot pretend this isn’t happening, and should not try to.

The evidence for antisickling plant compounds

The scientific record on traditional antisickling plants is more substantive than most conventional practitioners acknowledge. Cajanus cajan, pigeon pea, has demonstrated antisickling activity in vitro and has been the subject of multiple studies documenting its ability to reduce red blood cell sickling under deoxygenation. Zanthoxylum zanthoxyloides — fagara, well-known in West African traditional medicine — has shown meaningful antisickling properties in laboratory and animal studies, with Nigerian doctors documenting its clinical use for decades. A review tracing traditional sickle cell remedies in West Africa found that the most investigated plants — fagara, Carica papaya leaf, and several others — have pharmacological mechanisms plausible enough to justify continued study, and that the most important formal success story in this space, NIPRISAN, was developed directly from a Yoruba traditional formula and underwent a double-blind randomized placebo-controlled trial that demonstrated clinical benefit (Wambebe C et al., Journal of Ethnopharmacology, 2001, cited in Ojewole et al., Methods and Findings in Experimental and Clinical Pharmacology, 2006; Imoru et al., PMC, 2012).

NIPRISAN — a standardized extract of four plants (Piper guineense, Pterocarpus osun, Eugenia caryophyllata, and Sorghum bicolor) — became the first traditionally-derived herbal medicine to receive regulatory approval from Nigeria’s National Agency for Food and Drug Administration and Control. It has since encountered challenges around manufacturing consistency and commercial development, but its existence establishes an important principle: the knowledge embedded in West African traditional medicine is not merely anecdotal. Some of it is pharmacologically substantiated and clinically testable.

A 2019 ethnopharmacological study in the West Region of Cameroon documented 21 plant species used by traditional healers specifically for sickle cell disease management, and noted that several — including Cajanus cajan and Zanthoxylum zanthoxyloides — had existing laboratory evidence supporting their antisickling activity (Tagne RF et al., Evidence-Based Complementary and Alternative Medicine, 2022;9064533).

The evidence for herb-drug interactions: what pharmacists need to know

Respect for traditional medicine does not mean ignoring its risks. For patients concurrently using herbal preparations and prescription medications — which, in ATAcare’s patient populations, is common rather than exceptional — the interaction profile matters clinically.

The best-documented herb-drug interaction pathways in West Africa involve the cytochrome P450 enzyme family, the same enzyme system discussed in Post 2 in the context of pharmacogenomic variation. Several plants widely used in the region as antimalarials, cardiovascular remedies, and diabetes treatments have been shown to inhibit or induce CYP enzymes in ways that can meaningfully alter how prescription drugs are metabolized. Phyllanthus amarus and Tithonia diversifolia, popular in West Africa for malaria treatment, have been shown to inhibit CYP enzymes and the P-glycoprotein efflux transporter — creating the potential for elevated plasma concentrations of drugs that share those metabolic pathways, including some antiretrovirals and cardiovascular medications (Coulibaly SO et al., PMC, 2018).

For patients on warfarin — an anticoagulant used in cardiovascular disease and after mechanical heart valve replacement, a situation directly relevant to ATAcare’s population given the number of patients with rheumatic heart disease in our West African communities — Allium sativum (garlic), widely used as a cardiovascular remedy, has been documented to enhance the pharmacological effect of anticoagulants, increasing bleeding risk. The same plant has been shown to reduce the efficacy of certain antiretrovirals at high doses.

For patients on metformin, the first-line medication for type 2 diabetes, the picture is more nuanced: co-administration of Allium sativum with metformin has in some studies shown improved glycemic control, suggesting a potentially synergistic rather than antagonistic interaction. These are not simple, all-bad-or-all-good profiles. They require patient-level assessment — which is precisely what a pharmacist at a community clinic is trained to provide and is uniquely positioned to do.

The most important clinical reality here is not that traditional medicine is dangerous. It is that concurrent use of herbal and prescription medicines, when undisclosed, leaves the clinician without the full picture they need to assess drug response, side effects, and treatment failure. Patients who believe their healthcare provider will dismiss or disapprove of their traditional medicine use — which, based on the literature, many patients do believe — have every rational incentive to not disclose it. The result is incomplete clinical information that can lead to incorrect conclusions about medication efficacy, dosing adequacy, and patient adherence.

What a pharmacist-led organization is positioned to do

ATAcare’s approach to traditional medicine is built on a specific premise: that dismissing it is both clinically counterproductive and culturally disrespectful, and that engaging with it honestly is both more effective and more ethical.

Non-judgmental disclosure in medication counseling. Every medication counseling session at ATAcare’s clinic-pharmacy will include routine, non-judgmental inquiry about traditional and herbal medicine use. Not “are you taking anything you shouldn’t be?” but “many of our patients use both traditional remedies and prescription medications — can you tell me what you’re currently taking, including any plants or herbal preparations? This helps me make sure everything works together safely.” This single practice — normalizing disclosure — changes the clinical information available to ATAcare’s pharmacists and the providers they support.

Patient education on specific interactions relevant to our disease focus. ATAcare will develop plain-language patient education materials in French and English identifying the most clinically significant herb-drug interaction risks for patients on hydroxyurea, warfarin, antihypertensives, and antidiabetics — the drug classes most commonly used by the patients we serve. These materials will acknowledge the evidence for traditional medicine where it exists, note the interaction risks where they are documented, and avoid the dismissive framing that drives non-disclosure.

Building relationships with traditional healers as community health partners. In West African communities — and in many immigrant communities in the South Bronx and Upper Manhattan — traditional healers are trusted figures who often see patients before, during, and alongside their conventional care. ATAcare intends to pursue structured dialogue with traditional healers in Ouagadougou as community partners, not as competitors. The WHO’s Traditional Medicine Strategy 2019–2025 frames this integration explicitly as a goal of health systems globally, and several African countries have formal policies on traditional medicine integration. Burkina Faso has a documented national framework for this, however unevenly implemented.

Tracking the emerging pharmacological evidence. The scientific investigation of West African antisickling plants is ongoing, and the pharmacological logic behind some of these remedies is being studied with increasing rigor. ATAcare’s research mission includes following this literature and communicating what it finds to the families we serve — both when evidence supports a traditional remedy’s safety and efficacy, and when it identifies meaningful risks.

A note on what this series has documented

This post is the sixth in ATAcare’s AI & Chronic Disease Equity Series. Looking back across the series, a single through-line connects every installment: the patients ATAcare serves — in Upper Manhattan, the South Bronx, and Ouagadougou — have been systematically excluded from the infrastructure built to help them. Excluded from the training data that taught the algorithm. Excluded from the clinical trials that dosed the medication. Excluded from the validation studies that certified the sensor. Beyond reach of the gene therapy they most need. Moved through a care system that forgets them between visits. And sometimes — often — reaching for knowledge and care that their communities have always carried, in the form of plants and healers, because the formal system was not reliably there.

ATAcare is not large enough to fix any of this alone. But it is built specifically to operate in the space where these exclusions converge — at the intersection of pharmacy, research, and community health, in two cities and one country where that intersection is urgently needed.

• Tagne RF, Telefo PB, Nyemb JN et al. Ethnopharmacological study of medicinal plants used in the treatment of sickle cell anemia in the West Region of Cameroon. Evidence-Based Complementary and Alternative Medicine. 2022;9064533. doi:10.1155/2022/9064533
• Imoru M, Eke C, Ameh SJ, Onoja SO. Traditional herbal management of sickle cell anemia: lessons from Nigeria. International Journal of Herbal Medicine / PMC. 2012. PMC3502758.
• Wambebe C, Khamofu H, Momoh JA et al. Double-blind, placebo-controlled, randomised cross-over clinical trial of NIPRISAN in patients with sickle cell disorder. Phytomedicine. 2001;8(4):252–261. doi:10.1078/0944-7113-00045
• Herbal drug use in sickle cell disease management: trends and perspectives in sub-Saharan Africa — a systematic review. Current Drug Discovery Technologies. Eureka Select, 2019.
• Coulibaly SO et al. Modulation of Cytochrome P450, P-glycoprotein and Pregnane X Receptor by selected antimalarial herbs — implication for herb-drug interaction. PMC. 2018. doi:10.3390/molecules23102433
• World Health Organization. WHO Traditional Medicine Strategy 2019–2025. Geneva: WHO; 2019.
• Ismaila Shittu Abdullahi. Antisickling potential of aqueous extracts of Cajanus cajan leaf and seed, Zanthoxylum zanthoxyloides leaf, and Carica papaya (referenced in Imoru et al., 2012, and Tagne et al., 2022).