Discussion
This study describes the clinical presentation, microbiological diagnoses and visual outcomes of MK in the Terai region of Nepal, alongside patient interviews highlighting the barriers faced in accessing timely and appropriate clinical care. Patients affected by MK in this region tend to be of working age and engaged in agricultural work. This is a common trend among other studies of MK in LMICs including Uganda,6 Nepal7 and India.8 9 Two-thirds of patients were blind in the affected eye on presentation, and almost a half remained blind on their last outpatient visit. Poor visual outcomes from MK in LMICs are common; in Uganda 30% of patients had monocular blindness (visual acuity <3/60) at discharge,6 and in Tanzania 66% of patients had VA <6/60 at discharge.10
The majority of patients had seen another healthcare professional before attending the tertiary hospital (mostly ophthalmic assistant/optometrist or local pharmacy) and most were taking antibiotic eye-drops on arrival. Previous studies have shown that prompt use of prophylactic ocular antibiotics following a traumatic corneal abrasion can prevent MK from developing and results in much better outcomes.11–13 Despite the apparent common use of ocular antibiotics prior to patients presenting at the corneal department, visual outcomes remain poor. This suggests perhaps antibiotics were started too late after MK had already developed rather than being used prophylactically. Treatment for MK needs to be started as early as possible in order to achieve good visual outcomes, and once ulceration is advanced treatment is often ineffective.14 Of concern is that 21% of patients were taking corticosteroid eye-drops prior to presentation which can cause worsening of the infective corneal infiltrate, something that has been reported elsewhere.9
One predictor of a poor visual outcome at final follow-up was found to be trauma with vegetative matter. Other studies in Nepal and elsewhere in South Asia have also found that MK is commonly preceded by ocular trauma, with fungal keratitis more likely after trauma with vegetative matter leading to worse visual outcomes.11 15 16 More than half of all microbiological diagnoses were fungal, in keeping with other studies showing that fungal keratitis accounts for 20%–60% of MK in tropical regions.17 Fungal keratitis is challenging to treat and often requires a long and intensive treatment regimen which is difficult to adhere to due to the often limited availability and higher cost of the treatment. Even when promptly and adequately treated, up to 30% of infections still lead to corneal perforations or eye loss.18 19Brown et al found in their systematic review of literature on fungal keratitis that approximately 100 000 eyes are removed annually from fungal keratitis due to late diagnosis and poor therapeutic outcomes.20
The second significant risk factor for a poor visual outcome was delayed presentation from the start of symptoms to the eye hospital, with patients delaying between 15 and 30 days having a three-time increased risk of a poor visual outcome compared with those presenting within the first 5 days (OR 3.37, p=0.02 95% CI 1.19 to 9.50). The reasons for delayed presentation collected in the quantitative data strongly correlated with themes emerging from the qualitative data. Patient interviews provided greater insight into the reasons for delayed presentation to the eye hospital. Four common themes for delayed presentation emerged in the patient interviews which in turn can be used to plan methods to improve the patient journey, clinical experience and improve outcomes.
First, lack of awareness about the seriousness of this condition was common, with patients often not realising the importance of prompt treatment to prevent blindness. Patients highlighted the need for educational events in the community to raise awareness of the importance of accessing prompt treatment for eye trauma and ocular infections and where to find treatment.
Second, patients stated both a lack of available local eye services and a lack of knowledge about what local services were available. Nepal has a well-developed network of community eye care centres that feed into large regional eye hospitals and is fairly unique among LMICs for this organised structure of eye care services. The eye care centres are staffed by ophthalmic assistants, optometrists and eye health workers who have been trained to provide diagnoses, treatment and referrals for common ocular conditions. In Sagarmatha zone, where SCEH is situated, there are seven community eye care centres providing services to approximately 2.06 million people.21 As such, much of the population still have a long distance to travel to reach one of these centres, particularly in the hilly regions where there are fewer centres and travel is more challenging. Again, public health campaigns designed by the local eye care centres could help to raise awareness of their services and encourage patients to attend as well as possible long-term plans to increase the number of eye care centres in areas further away from SCEH.
Third, while the majority of patients had a prior consultation in the community before attending SCEH, many stated in their interviews that this consultation had led to delay, incorrect treatment and false reassurance. This finding is supported by a recent study of the knowledge of MK among primary healthcare workers in the same region. We found that only 27% of primary healthcare workers in community health posts could correctly diagnose MK and 59% could correctly treat the condition. Only 69% were aware that it is an ocular emergency requiring prompt treatment and referral.22 This study supports the findings from Burn et al that primary eye care training is required in the region, particularly focused on common blinding conditions such as MK, in order to avoid delays in referral and harmful medical treatments being offered (such as traditional eye medicines (TEM) and corticosteroid eye-drops).
Lastly, patients described costs, lack of family support and nobody to accompany them as important barriers to accessing eye care. In particular, the medication prescribed to treat fungal keratitis is often expensive and not locally available meaning that patient compliance with treatment can be poor. Improving the availability and affordability of anti-fungal medication, in particular, is a key priority in improving outcomes from MK in this region. A current randomised control trial is investigating the use of a cheap and more widely available medication, chlorhexidine 0.2% eye-drop, as a non-inferior alternative to natamycin 5% eye-drop which is currently first line.23 Although on the WHO Essential Medicines List 2017, natamycin 5% eye-drops are often not available in much of sub-Saharan Africa and some Asian countries, and where it is available it is prohibitively expensive.10
The main limitation of this study is the potential for selection bias and researcher bias in the qualitative interviews. While we aimed to maintain a neutral position while carrying out the interviews, the personal beliefs and experiences of the interviewer can still influence the response of the participant. As purposive sampling was carried out those participating in the interviews may not be fully representative of participants attending the corneal department at SCEH. Second, corneal samples were not cultured as this facility was not available at the hospital. Culture of corneal scrapes is the gold standard method for identifying bacterial corneal infections. Without this diagnostic tool the microbiological diagnoses have been made only from Gram stain and microscopy limiting the accuracy of the results and could account for a low bacterial detection rate. Lastly, this study only collected data at one tertiary eye hospital in the region and would be improved by expanding the sample to include multiple sites to allow greater generalisability of the results.
