Smart Cards in Malawi Matt Boxshall The Lighthouse

Smart. Cards in Malawi Matt Boxshall The Lighthouse Trust Lilongwe, Malawi mattb@sdnp. org. mw

Malawi l l l Pop 11 m GDP / Capita (2000) = US$170 65% ‘poor’ - unable to meet daily nutritional needs (NSO 2000) HIV prevalence - Urban 22. 5%, Rural 10. 7%, around 1 million infected Life expectancy dropping, < 40 yrs Pop / nurse approximately 3, 500, or about 1 per 100 HAART eligible patients

HAART in Malawi Approximately 3, 000 registered on HAART, mid 2003 l Four ‘formal’ sites currently operational, total capacity to register perhaps 3500 new HAART clients annually l Global Fund money will pay for free HAART for >>25, 000 over 5 years l

The Lighthouse l Background - Hospital Volunteers, Complimentary services, Trust working as a PPP l Strategy - Scale, Model, Build Capacity l Services - CHBC, VCT, Clinic - HAART

Cumulative HAART Registrations Graph Reg l HAART registration graph LCH to Lighthouse • Government Drugs @ US$ 30 / month • Demand vs Supply

Lighthouse Clinic Monthly Client Visits Graph Clinic visits • Cumulative Workload • Reaching Capacity - where to next?

Response “Fast Track” to move review non-problematic reviews to more junior staff (nurses) l Decentralize reviews to health centers l l BUT patient data management systems are also increasingly stretched, and decentralization (and ARV shopping) will only exacerbate this: ØHow do we identify and follow our patients? ØHow do we know who fails to pick up their drugs (or picks up more than one supply? ) ØHow do we gather information centrally for M+E? ØHow do we account for drugs?

A Technological Fix? We can’t throw people at this problem - we don’t have them! l We need something; l – Easy to use – Robust – Scalable – Tamper-proof – Reasonably priced

Smart. Cards A programmable chip on a credit card, read by a “point of sale device” (Po. S) l Successfully implemented in the region KWS, petrol stations, banks, micro-finance projects etc l Local providers available and interested (Malswitch, NET 1) l Costs approx $5 / card, rental of Po. S approx $25 / month - small vs drug costs l

Programming the Card l Cards issued at prescribing site l Each card has 550 fields or ‘wallets’ l Fields can be entered at registration, updated at drug collection, calculated, password protected etc. l Biometric information can be carried - in this case, fingerprint scans

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Sample Fields Fingerprint Biometrics x 2 Patient ID number Date of start of ART Date of Registration onto Smart. Card ART Registering Clinic Name Registering Clinician Name Date of first collection of drugs with Smart. Card Drug regimen details (+ change regimen flag? ) Date of last drug collection Date current drug supply will finish Location of last drug collection Name of person dispensing drugs Number of pills dispensed Collection by Patient or Guardian Cumulative Total Pills received Patient Working Drug Credit Default Flag

Drug Collection l l l l Any patient should be able to pick up drugs anywhere a Po. S device is available Patient (or Guardian) identified offline Automated checks run (eg late collection) Drug collection authorised, details updated to card Details downloaded to Po. S Vendor card updated Printout if required

Data Collection Patient data collected electronically and largely automatically at Po. S l Po. S downloaded either by dial-up, or by transfer to ‘milking’ card l Drugs credited to Cards for transfer between sites, and stock management (at least partly) automated l Drug and Patient management, M&E, closely linked l

Unresolved Issues? l l Health worker uptake will be dependant on perceived value, particularly in time saved Patients may resist, preferring less ‘control’ (although balanced vs flexibility of collection site) Centralized electronic data collection may raise confidentiality issues Responsibility for system management may be divisive - clinical services, medical stores?

The Way Forward l l Technical specifications have been drafted with suppliers Lighthouse will initiate with partners in Mo. HP (and others) Operational research should evaluate effectiveness If successful, roll-out will need to be fast to establish system in line with planned HAART scale-up - system makes a lot more sense if it is country wide

Acknowledgements l Lighthouse - Sam Phiri, Florian Neuhann, Ralf Weigel l University of North Carolina - Mina Hosseinipour l Baobab Health Partnerships - Richard Altmann, Gerry Douglas l Net 1 - Brenda Stewart

BAOBAB