Findings from this pilot study show that knowledge about eye conditions in children and how they can be prevented or managed was low among RCH staff in clinics in Dar-es-Salaam, even for common conditions such as conjunctivitis. Indeed, some staff suggested inappropriate treatments, such as steroid drops for conjunctivitis or vitamin A capsules for cataract. This inappropriate treatments are likely to exacerbate the condition, or lead to delay in accessing appropriate care, which in the case of cataract, increases the risk of poor visual outcomes
[9, 10, 16]. The relatively low level of knowledge almost certainly reflects the limited exposure staff would have had to primary eye care during their initial training. Also as noted in the results, preventing ophthalmia neonatorum is the only eye-specific condition included in the initial curriculum. The lack of educational materials on eye conditions in the clinics also suggests that eye care had not been a topic covered in training or continuing professional development. Our findings are in contrast to a study in Ethiopia where primary health care workers were more knowledgeable about the causes of blindness in children
 probably because of the on the job training they had received.
Staff in the study were eager to learn and training improved their knowledge and management decisions, as in a study in India
. It is important to note that knowledge was better at three weeks than at one year after training. This emphasizes the need for supportive supervision, which should focus on continuing professional development and not just administrative or managerial matters
. Studies elsewhere have shown that training alone can give disappointing results in terms of changing practices and behaviours
. A future study will need to explore barriers which may limit the ability of RCH staff to change their practices e.g. lack of autonomy in decision making, unclear roles and responsibilities, or lack of time, so that these barriers can be addressed
In our study COs were generally more knowledgeable than nurses, and could name more eye conditions and their management than nurses before training, reflecting their different educational backgrounds and roles. However, some COs recommended some potentially harmful treatments, such as treating conjunctivitis with steroids. Although COs diagnose and treat conditions, nurses are the first point of contact with mothers and their children, and provide the health education. Although the study did not explore the satisfaction of RCH workers with the training methods, duration and training materials in any depth, some participants thought the training was too short. Future studies should explore the optimal duration of training for different cadres of RCH staff considering that they have different educational backgrounds. Most staff were enthusiastic about the training materials but any further studies should explore whether they can be improved.
Practice of RCH staff
Staff were already providing a good service in relation to measles immunization and vitamin supplementation but they did not know that these interventions also reduce blindness in children. The high coverage of these interventions, which reflects the priorities and policies of the Ministry of Health, are reducing under 5 mortality rates in Tanzania. There is also anecdotal evidence of less corneal blindness in children
. However, activities specific to eye health, such as Crede’s prophylaxis, were not being routinely performed. Reasons for not using Crede’s prophylaxis included overwork, and lack of supervision and eye drops. However, after training half of the facilities that conducted deliveries resumed the practice, which suggests that improved knowledge in relation to the importance of ophthalmia neonatorum motivated staff to change their practices, which were maintained at one year. The number of children referred to MM Health centre increased after training, but a limitation of this study was that it was not possible to track these children to ascertain how many were referred unnecessarily, nor to determine the accuracy of diagnoses made by RCH staff. Efforts to track referred patients and their diagnoses should be made in a future study.
Another positive change was that trained staff felt confident to deliver health education on eye conditions, often including eyes when talking about other conditions, such as measles. However, to be effective, health education should take place alongside improvement in the quality of services, in both clinical and non clinical care
. Topical medication and diagnostic equipment, such as torches, must be provided, to enable staff to put into practice what they talk about during health education.
Educational materials for those using health services are a requirement of PEC
 and the lack of information on eye conditions for mothers as well as staff shows that this aspect of PEC also needs to be addressed. In this study efforts were made to make the poster relevant and attractive, with messages that were easy to understand. Staff reported that mothers responded positively to the poster and they valued and used the poster and manual for reference and health education.
The misconceptions and beliefs in the community about eye conditions reported by staff in this study may partly explain delay in accessing eye services for children. Other studies in Tanzania and Sub-Saharan Africa show that home treatments and remedies, and those recommended by traditional healers, are commonly used for eye conditions in children
. Some of these practices can be harmful (e.g., causing trauma, infection) while others can lead to delay in seeking appropriate treatment. It is important that staff in RCH clinics address these topics during health education