Data mobility, not algorithms, will fix Nigeria’s health system- Ilodigwe

Lucky Ilodigwe is a healthcare consultant at Boston Consulting Group (BCG) in the United States. He previously served as a healthcare strategist at PwC Nigeria, where he advised public-sector health insurance schemes and tertiary institutions. A medical biochemist by training, he holds an MBA from the University of Chicago Booth School of Business, where his work focused on healthcare strategy and the economics of health systems in emerging markets. In an exclusive interview with Pharmanewsonline, the Nigerian-born healthcare strategist, now advising United States payers and providers at Boston Consulting Group, argues that the country’s path to AI-enabled care runs not through imported algorithms, but through a patient identifier, interoperable records, and the institutional discipline to use the data the system already generates.

“Nigeria is not short of healthcare data,” Ilodigwe told Pharmanewsonline in an extended conversation. “We are short of healthcare data that moves.”

According to the National Bureau of Statistics, out-of-pocket spending still accounts for roughly 70 to 75 per cent of total health expenditure in Nigeria, one of the highest ratios in the world. Every naira of that spend leaves a data trail somewhere: a receipt at a community pharmacy, a card at a Primary Health Centre, an HMO claim, a hospital ledger. Almost none of it is connected. The cost of that disconnection, Ilodigwe argues, is paid by patients, in repeated tests, delayed diagnoses, and avoidable hospitalisations.

He is also blunt about how the global AI conversation is reaching Nigeria.

“The dominant framing is imported largely from American and European headlines, AI radiologists outperforming consultants, large language models passing medical licensure exams, generative tools drafting clinical notes,” he said. “These are real advances, and worth following. But they are not where the leverage lies for a system like ours. The leverage for Nigeria lies upstream of the algorithms, in the unglamorous work of making data usable in the first place.”

The real problem is not intelligence. It is plumbing.

In his time advising health systems and payers, first at PwC in Lagos, where he supported state-level health insurance schemes and federal tertiary institutions, and more recently at BCG, working across United States payers and providers, Ilodigwe says one pattern has been remarkably consistent across geographies.

“The systems that get the most out of analytics and AI are not the ones with the most sophisticated models,” he said. “They are the ones that did the boring work first.”

That boring work, as he describes it, has four components. Nigeria, he assesses, is at varying stages of maturity on each.

The first is a single patient identifier. “Without a way to recognise that the Tunde Adeyemi who visited a PHC in Surulere in March is the same Tunde Adeyemi admitted to LUTH in August, every other data initiative breaks down,” he said. He points to the National Identity Management Commission (NIMC), which has now enrolled over 100 million Nigerians for the National Identification Number (NIN). Linking the NIN to a national health record, as Estonia did with its e-ID and as India is now doing through the Ayushman Bharat Digital Mission, is, in his view, the single highest-leverage data investment Nigeria could make this decade. He sees the National Health Insurance Authority (NHIA), reconstituted under the 2022 NHIA Act, as the natural anchor.

The second is interoperable clinical records. Even within a single Lagos State General Hospital network, electronic medical record (EMR) adoption is uneven. Where EMRs do exist, Helium Health and Wellvis have made meaningful inroads in the private and faith-based hospital segment, systems rarely speak to one another.

“The FHIR standard is not exotic technology. It is a published specification,” Ilodigwe said, referring to Fast Healthcare Interoperability Resources, the global benchmark for clinical data exchange. “Mandating FHIR-conformant data exchange for any facility participating in NHIA, the formal sector schemes, or the new Basic Healthcare Provision Fund would, within three years, create the connective tissue that today does not exist.”

The third is claims and pharmacy data analysed at the population level. State health insurance agencies, Lagos State Health Management Agency (LASHMA), and the schemes in Kwara, Ogun, Edo, and other states stood up under the Act, sit on transaction-level data that, properly de-identified and pooled, could illuminate disease burden, prescribing patterns, and fraud in ways the system has never been able to see. mPharma’s pharmacy network and the major HMOs hold complementary data. “None of it is currently being analysed at the scale the underlying volume would permit,” he said.

The fourth, and in his view the most often overlooked, is data governance. “Who owns a patient’s record? Under what conditions can it be used for secondary analysis? How is consent managed?” he asked. The Nigeria Data Protection Act of 2023 created the legal scaffolding, he noted. The Nigeria Data Protection Commission now needs to issue health-sector-specific guidance, and health system leaders need to invest in the institutional capacity to comply with it. “This is not a constraint on AI,” he said. “It is a precondition for it.”

Five use cases that fit Nigeria’s reality

Once that plumbing exists, Ilodigwe says, the application layer becomes interesting. He is sceptical that the most useful AI in Nigerian healthcare over the next five years will be a foundation model trained in San Francisco. “It will be narrow, embedded, and matched to bottlenecks our clinicians and administrators already know exist,” he said. He offers five examples he considers immediately tractable.

 Maternal risk stratification at the PHC level. Nigeria still accounts for roughly 28.5 per cent of global maternal deaths, according to the most recent WHO estimates, the highest absolute burden of any country. The clinical risk factors for postpartum haemorrhage, eclampsia, and obstructed labour are well established. A lightweight model running on the tablet a community health extension worker already carries, drawing on antenatal visit history and a handful of vitals, could flag the women in a catchment area who most need to be transferred to a comprehensive emergency obstetric care facility before labour begins. “The National Primary Health Care Development Agency’s revised PHC platform already envisions digital tools at the ward level,” Ilodigwe noted. “The model is the easy part. The referral pathway and the ambulance, those are the hard parts.”

 Tuberculosis and oncology imaging triage. Chest X-ray AI for TB screening, qXR by Qure.ai and CAD4TB by Delft Imaging, is already endorsed by the World Health Organization and deployed in several African countries. Nigeria has the world’s sixth-highest TB burden and a radiologist density of fewer than one per 100,000 people. “A triage algorithm at LGA-level diagnostic centres, reviewed by a teleradiologist, is a more realistic path to closing the diagnostic gap than waiting to train enough radiologists to cover 774 local government areas,” Ilodigwe said. The same logic, he argues, applies to cervical cancer screening via automated visual inspection, where the technology is mature.

Drug supply and stockout prediction. Anyone who has worked in Nigerian public health knows the rhythm: a state runs out of antiretrovirals, malaria commodities, or oxytocin, and the consequence is measured in lives. Ilodigwe cites LifeBank as proof of what a coordinated logistics layer can do for blood. Demand forecasting on consumption data from the Nigeria Supply Chain Integration Project, combined with the Federal Ministry of Health’s logistics management information system, is, in his words, “a tractable AI problem, and one where a 10 to 15 per cent reduction in stockouts produces measurable, attributable health gains

HMO and insurance fraud and leakage detection. Healthcare fraud in Nigeria is not a marginal issue. Industry estimates from the NHIA and private HMOs put leakage, phantom enrollees, upcoding, duplicate claims, ghost visits, at meaningful double-digit percentages of payouts. “This is precisely the kind of pattern-recognition problem at which machine learning excels,” Ilodigwe said. “And it is the use case most likely to generate the savings that fund the rest of the digital agenda.”

 Chronic disease registries and longitudinal care. Hypertension affects roughly one in three Nigerian adults; diabetes prevalence is climbing past 5 per cent and rising fastest in urban centres. “These conditions are managed over decades, not encounters,” Ilodigwe said. A linked registry, even a regional one, built on top of an NHIA-anchored identifier, would, in principle, allow a cardiologist in Port Harcourt to see whether her patient filled his amlodipine prescription in Aba last month. “Today she cannot. This is not a technology gap. It is an integration gap.”

What the workforce conversation gets wrong

The most consequential health workforce story in Nigeria today is japa, the steady departure of doctors, nurses, and pharmacists to the United Kingdom, the United States, Canada, and increasingly the Gulf states. The Medical and Dental Council of Nigeria has documented the licensure verification requests; the United Kingdom’s General Medical Council has registered the inflow. By every credible estimate, Nigeria now produces health professionals at a slower rate than it loses them.

It is tempting, in this context, to position AI as a substitute. Ilodigwe rejects that framing.

“Clinicians can smell that argument from a long way off, and they are right to,” he said. “The honest case is that AI and well-designed digital tools allow the clinicians who remain to operate at the top of their training, to spend less time hunting for results, transcribing notes, and chasing referrals, and more time on the judgment that only they can provide. Ambient documentation tools, clinical decision support embedded in EMR workflows, asynchronous specialist consultation platforms, these are not replacements for the consultant who emigrated. They are how the consultant who stayed survives a clinic of 80 patients.”

The same logic, he says, applies to community health workers, who carry far more of Nigeria’s actual care delivery than the formal physician numbers suggest. “The right software in their hands extends their reach. The wrong software, extractive, top-down, designed in a conference room in Abuja or Geneva, adds reporting burden and nothing else,” he said. “Which one gets built is a leadership choice, not a technology question.”

Governance, trust, and the cost of getting this wrong

Two risks deserve to be named plainly, Ilodigwe says, because the international discourse on AI in healthcare often skirts them.

The first is algorithmic bias. Almost every clinical AI tool now in commercial use was trained predominantly on data from North American, European, or East Asian populations. The performance of these tools on Nigerian patients, different disease prevalences, different presentations, different comorbidity profiles, frequently darker skin tones for dermatological models, cannot be assumed.

“NAFDAC’s existing remit covers medical devices,” Ilodigwe said. “The agency, alongside the Federal Ministry of Health and the Nigeria Data Protection Commission, will need to define what local validation looks like before a foreign-trained model is approved for clinical use at scale. This is not protectionism. It is patient safety.”

The second is the trust deficit. Nigerians, he observes, have good reason to be cautious about how institutions handle their personal information.

“A health AI strategy that does not begin with explicit, plain-language patient consent, transparent data-use rules, and meaningful redress when things go wrong will not survive its first public failure,” he warned. “The Lagos State Health Scheme, the NHIA, and the major HMOs have an opportunity to set the standard here. They also have a responsibility to.”

A pragmatic agenda

Asked what he would tell the Federal Ministry of Health, the NHIA, or a forward-leaning state government considering where to start, Ilodigwe outlines a deliberately narrow agenda.

Anchor a national patient identifier to the NIN within the next eighteen months, with the NHIA as the convening institution. Mandate FHIR-conformant data exchange as a condition of participation in any publicly funded scheme. Stand up a federal-level health data analytics unit, small, technical, insulated from political churn, with the mandate to analyse claims and clinical data and publish what it finds. Fund three to five demonstration projects in domains where the use case is mature: TB imaging triage, maternal risk stratification, and commodity stockout forecasting, in his view, are the obvious candidates. And invest, seriously, in the Nigeria Data Protection Commission’s capacity to regulate health-sector data with the rigour the moment requires.

“None of this requires a breakthrough,” he said. “It requires sequencing, discipline, and the willingness to do the unglamorous work before reaching for the shiny tool.”

He returns, in closing, to a comparison that he believes Nigerian policymakers should sit with.

“The countries that will get the most out of artificial intelligence in healthcare over the next decade are not the ones with the most advanced models. They are the ones whose health systems are organised enough to use them,” Ilodigwe said. “Nigeria has every ingredient it needs to be in that group, the talent, the data, the policy frameworks, the entrepreneurial energy in companies like Helium Health, mPharma, 54gene, and LifeBank. What it needs now is the institutional patience to build the foundation before the building.”

“We have spent two decades expanding access to care, with real if uneven gains,” he added. “The next decade will be defined by whether we can make that care intelligent. The tools to do it exist. The question is whether we choose to use them well.”