While access to primary health care services has improved through expansions of the public and private sector, quality of care remains a pressing issue in the Middle East and North Africa region. This thesis focuses on Tunisia, where there have been reports of challenges in the availability of qualified health workers and essential medicines. However, there is less evidence on the clinical quality of care, which is more difficult to monitor and regulate. This thesis investigates the quality of care provided by general practitioners (GPs) and its determinants in Tunisia. To measure quality, I use undercover standardised patient visits and clinical vignettes carried out between November 2017 and June 2018 from a representative sample of 177 public and 79 private GPs in the urban ‘Grand Tunis’ region of Tunisia. These data were combined with information from a GP survey data on provider characteristics. The thesis includes three research papers. In the first one, I evaluate doctor clinical knowledge and investigate its determinants, with a key focus on differences between public and private doctors. I find that clinical knowledge was suboptimal, with wide variation across clinical cases and lower knowledge in the private sector. The place of initial medical education was a strong predictor of clinical knowledge. In the second paper, I examine antibiotic prescribing behaviour and the role of the patient through an experimental audit study involving standardised patients. Overprescribing rates were strikingly high, with prescribing behaviour significantly worse in the private sector. The majority of the drugs prescribed were unnecessary or potentially harmful and generated high levels of waste of patients’ resources. There was limited evidence that provider behaviour was influenced by standardised patients requesting antibiotics or expressing knowledge of inappropriate prescribing. In the third paper, I explore whether doctors discriminate according to patient socioeconomic status (SES). I find no evidence of SES-based discrimination in the provision of care to standardised patients. If anything, I find that doctors prescribed less expensive medicines and gave significantly more (and more unnecessary) free-of-charge medicines to poorer patient. However, doctor communication was worse when patients were poor than when they appeared well off. Overall, these results suggest that doctors play a crucial role in driving the quality of primary health care in urban Tunisia. This quality of care appears to vary in some respects with the socioeconomic background of the patient. Further, there is widespread overprovision of antibiotics in both the public and private sector, but doctors’ behaviour does not appear to be influenced by the patient. With these findings, I consider the policy implications and future avenues for further research.