Background Novartis Access is a Novartis programme that offers a portfolio of non-communicable disease medicines at a wholesale price of US$1 per treatment per month in low-income and middle-income countries. We evaluated the effect of Novartis Access in Kenya, the first country to receive the programme. Methods We did a cluster-randomised controlled trial in eight counties in Kenya. Counties (clusters) were randomly assigned to the intervention or the control group with a covariate-constrained randomisation procedure that maximised balance on a set of demographic and health variables. In intervention counties, public and non-profit health facilities were allowed to purchase Novartis Access medicines from the Mission for Essential Drugs and Supplies (MEDS). Data were collected from all facilities served by MEDS and a sample of households in study counties. Households were eligible if they had at least one adult patient who had been diagnosed and prescribed medicines for one of the non-communicable diseases targeted by the programme: hypertension, heart failure, dyslipidaemia, type 2 diabetes, asthma, or breast cancer. Primary outcomes were availability and price of portfolio medicines at health facilities, irrespective of brand; and availability of medicines at patient households. Impacts were estimated with intention-to-treat analysis. This trial is registered with ClinicalTrials.gov (NCT02773095).

Findings On March 8, 2016, we randomly assigned eight clusters to intervention (four clusters; 74 health facilities; 342 patients) or control (four clusters; 63 health facilities; 297 patients). 69 intervention and 58 control health facilities, and 306 intervention and 265 control patients were evaluated after a 15 month intervention period (last visit February 28, 2018). Novartis Access significantly increased the availability of amlodipine (adjusted odds ratio [aOR] 2·84, 95% CI 1·10 to 7·37; p=0·031) and metformin (aOR 4·78, 95% CI 1·44 to 15·86; p=0·011) at health facilities, but did not affect the availability of portfolio medicines overall (adjusted β [aβ] 0·05, 95% CI –0·01 to 0·10; p=0·096) or their price (aβ 0·48, 95% CI –1·12 to 0·72; p=0·500). The programme did not affect medicine availability at patient households (aOR 0·83, 95% CI 0·44 to 1·57; p=0·569).

Interpretation Novartis Access had little effect in its first year in Kenya. Access programmes operate within complex health systems and reducing the wholesale price of medicines might not always or immediately translate to improved patient access. The evidence generated by this study will inform Novartis’s efforts to improve their programme going forward. The study also contributes to the public evidence base on strategies for improving access to medicines globally