Will improvement of knowledge lead to improvement of compliance with glaucoma medication?

Editor, We read with interest the paper by Hoevenaars et al. (2008). The authors are to be complimented for designing an extensive study attempting to find out whether addressing specific items of improved knowledge of glaucoma leads to better compliance. In their study, 44 randomly selected ophthalmologists each selected four consecutive patients (n = 166) to complete a structured questionnaire. The results could not show a statistically significant correlation between the total level of knowledge and compliance. The authors conclude that it would be ‘unlikely that further improving knowledge will greatly improve compliance with glaucoma medication’. They recommend that strategies on enhancing compliance should not focus on further improving knowledge and that a patient education programme will not be of added value in improving compliance. We beg to differ with their conclusions. Firstly, their counterintuitive findings are in contradiction to most of the published literature, which shows a positive correlation between better understanding of glaucoma and compliance (MacKean & Elkington 1983; Zimmerman & Zalta 1983; Khandekar et al. 2005). Recently, this relationship was shown to have a risk ratio (RR) of 2.05 [95% confidence interval (CI) 1.07–3.93] in a study of 105 Omani and a RR of 4.54 (95% CI 1.45–14.16) in 200 Swiss glaucoma patients (K. Mansouri et al., unpublished). However, because most studies on compliance lacked sufficient power to assess the effect of patient knowledge and compliance, a significant majority of participants did not have an accurate viewpoint of glaucoma, thus reducing the numbers available for analysis (Konstas et al. 2000). The educational level of patients is also positively correlated with better compliance (Sleath et al. 2006): several studies have demonstrated that less formally educated individuals, particularly those with less than secondary education, were more unlikely to be familiar with glaucoma (Gasch et al. 2000; Lau et al. 2002; Olthoff et al. 2005; Hoevenaars et al. 2006). Also, patients with chronic disease (including glaucoma) frequently complain of a lack of information (Osterberg & Blaschke 2005; Mansouri et al. 2006), making this a major barrier to improving compliance. Interestingly enough, Hoevenaars et al.’s recent recommendation is also in contradiction of an earlier report by the same group that analysed the same cohort for socioeconomic determinants, which suggested that ‘better provision of information to patients … could reduce their risk of becoming blind … and improve compliance’(Hoevenaars et al. 2006). Their study suffers from a number of limitations that might have contributed to the surprising findings: The definition of non-compliance (one missed dose in the past 4 weeks) is unusual. Most reports use a less strict criterion of one or more doses in the past week. The definition of ‘total level of knowledge’ (some questions were too complex for an average patient) does not necessarily reflect useful knowledge of glaucoma. The response rate was relatively low, with only 166 out of 240 targeted patients (69.1%) and 73% of ophthalmologists participating compared to the reported 85% of patients. Eleven patients did not fill in the number of doses missed and a further 12 were excluded for various reasons, reducing the effective response rate to 59.5%. It could further be argued that selecting only four patients per practice could lead to potential selection bias and chance. Furthermore, the issue of dyscompliance (how to correctly instill drops), an important part of achieving a treatment response, was not addressed in the study. It can be reasonably argued that dyscompliance could be reduced by adequate education and the provision of instillation aids through the healthcare provider. We believe that the study design does not allow the authors to arrive at the above-mentioned conclusions. The authors should have conducted an interventional trial with a study group receiving education on glaucoma and a control group not receiving any additional education. Even in the absence of these shortcomings, the conclusion drawn would still not be justified: the authors themselves demonstrate that three out of 22 items showed a significant positive correlation with compliance (such as patients knowing that eyedrops cannot repair glaucoma damage), and a further 19 items showed a positive relation with compliance while not being significant. Therefore, we would like to caution practitioners and policy-makers to interpret these conclusions carefully and in the context of the other published literature. Nevertheless, we appreciate the effort invested by the authors to address an often neglected field of glaucoma research.