Measurement of food literacy among the adult population in urban Uganda and Kenya: development and validation of an East African food literacy scale

Objective: Food literacy (FL) is a potential approach to address the nutrition transition in Africa, but a validated tool is lacking. We developed and validated a scale to assess FL among Ugandan and Kenyan adult populations. Design: A mixed-method approach was applied: (1) item development using literature, expert and target group insights, (2) independent country-specific validation (content, construct, criterion and concurrent) and (3) synchronisation of the two country-specific FL-scales. Construct validity was evaluated against the prime dietary quality score (PDQS) and healthy eating self-efficacy scale (HEWSE). Setting: Urban Uganda and Kenya. Participants: Two cross-sectional cross-country surveys, adults >18 years (n = 214) and university students (n = 163), were conducted. Results: The initial development yielded a forty-eight-item FL-scale draft. In total, twenty-six items were reframed to fit the country contexts. Six items differed content-wise across the two FL-scales and were dropped for a synchronised East African FL-scale. Weighted kappa tests revealed no deviations in individuals' FL when either the East African FL-scale or the country-specific FL-scales are used; 0<middle dot>86 (95 % CI: 0<middle dot>83, 0<middle dot>89), Uganda and 0<middle dot>86 (95 % CI: 0<middle dot>84, 0<middle dot>88), Kenya. The FL-scale showed good reliability (0<middle dot>71 (95 % CI: 0<middle dot>60, 0<middle dot>79), Uganda; 0<middle dot>78 (95 % CI: 0<middle dot>69, 0<middle dot>84), Kenya) and positively correlated with PDQS (r = 0<middle dot>29 P = 0<middle dot>003, Uganda; r = 0<middle dot>26 P < 0<middle dot>001, Kenya) and HEWSE (r = 0<middle dot>32 P < 0<middle dot>001, Uganda; r = 0<middle dot>23, P = 0<middle dot>017, Kenya). The FL-scale distinguishes populations with higher from those with lower FL (beta = 14<middle dot>54 (95 % CI: 10<middle dot>27, 18<middle dot>81), Uganda; beta = 18<middle dot>79 (95 % CI: 13<middle dot>92, 23<middle dot>68), Kenya). Conclusion: Provided culture-sensitive translation and adaptation are done, the scale may be used as a basis across East Africa.