Glaucoma treatment and deprivation: time-series analysis of general practice prescribing in England

Summary

Our study demonstrates PGAs accounted for approximately two-thirds of general practice prescriptions for glaucoma-related diagnoses in England over 6 years. After the first NICE guideline update in 2017, use of PGAs increased for the next 3 years to ~640 000 prescriptions per month, a ~10% increase. When analysing prescribing per 1000 population, there was an overall downward trend over the 6-year period, a finding potentially contrasting to the expected increase in at-risk cases needing treatment. In April 2020, there was a more rapid prescribing decline as COVID-19 impacted on sight testing, with subsequently reduced referrals and fewer new diagnoses. Significantly, in more deprived areas, crude prescribing per 1000 population was substantially lower than in more affluent areas, suggesting matters have not improved since Owen et al’s early study.⁹ Factors influencing this finding may include lower referral from primary care optometry within areas of deprivation, more advanced presenting glaucoma within deprived areas (leading to surgery as part of primary treatment or escalation to surgery after initial PGA prescribing), and poorer medication adherence. The latter may be linked to inadequate patient education and/or language provision, impacting on attendance and culminating in failure to refill prescriptions.

Strengths and limitations

A strength of this work is collation of large-scale glaucoma prescribing in the NHS in England. Arguably these data (>9 million prescriptions per annum) reflect an unbiased source. However, while these data provide a complete picture of primary care prescribing in England, secondary care prescribing data, even when community dispensed, were not captured. Further, the dataset was not a record level dataset, meaning it did not include demographic information, for example, age or ethnicity. In addition, the analysis relies on the GP practice base for deprivation, rather than providing a view of deprivation for individuals receiving prescriptions. The IMD score was updated in 2019 and deprivation scores are relatively up-to-date descriptors of deprivation status. We acknowledge that this 2019 version may be outdated. Furthermore, allocation of one IMD score per practice may not adequately reflect populations, which likely has patients with different IMD levels, an acknowledgement of the well-reported complexity of SES, neighbourhood deprivation, and health.¹⁶ However, despite limitations, IMD remains the best readily available method for examining deprivation, while accepting some areas may have seen improvements in SES, for example, 'gentrification' masking existing population requirements, causing apparent improvement in deprived areas.

A further limitation is that our analyses were restricted to PGA prescribing, and future analyses may usefully include all glaucoma medications. Furthermore, we concede the historic nature of the data. We were not able to identify individuals, duration of using PGAs, or glaucoma stage at diagnosis. People presenting with advanced glaucoma may be managed differently, with NICE recommending early surgery, hence overall PGA prescribing may be lower owing to surgical versus medical management; however, this explanation was not borne out by a study of SES in cases undergoing trabeculectomy,¹⁷ where drainage surgery was carried out less frequently in patients from areas of greatest deprivation. While it is uncertain whether surgical management impacts PGA prescribing in areas of deprivation, there is evidence patient adherence may be influential. A study looking at demographics of patients registered severely sight impaired from glaucoma¹⁸ found those with advanced disease and those with poor adherence were more likely to have greater deprivation. Arguably, the corollary is increased sight testing in affluent areas generates more glaucoma referrals and greater repeat prescribing within primary care. Further research is necessary to better understand these differences, although since initial prescribing is largely in secondary care (covering populations from the most and least deprived areas), with repeat primary care prescribing supported from secondary care, it seems likely differences in deprived area versus national data include fewer referrals, more advanced presenting disease, and prescription refill factors.

Comparison with existing literature

The majority of glaucoma cases are detected through case finding in primary care optometry. Day et al ¹⁹ showed a mismatch between areas of deprivation and location of optometrists, a finding supported by our own recent analysis,²⁰ with both studies supporting the view that the optometry business model may deter practice establishment within deprived communities,²¹ creating a barrier to sight testing, impacting on detection. Several studies have looked at access to eyecare and SES. Knight et al’s review²² noted seven of eight high-quality studies concluded there was a significant positive association between lower SES and glaucoma stage at presentation, and a significant negative association between SES and secondary care attendance. An early case-control study examining deprivation and stage of presentation²³ found deprivation was associated with late presentation, an important risk factor for subsequent blindness. A cohort study in Manchester investigating SES and vision loss in glaucoma²⁴ found patients from deprived areas presented with more advanced loss, while SES has been shown as a risk factor for patients with acute primary angle closure.²⁵ More recently, the relationship between late presentation of glaucoma and deprivation has been revisited, with Rathore et al ²⁶ confirming the association between IMD and advanced visual field loss at diagnosis, while concluding rapid worsening of glaucoma during follow-up was not associated with IMD, suggesting equity of care and outcomes once patients were referred into the English hospital eye service. This latter suggestion is supported by King et al,²⁷ who observed in their treatment of advanced glaucoma study that while SES at baseline is correlated with poorer vision it did not impact on the success of treatment at 24 months.

Some previous studies have described associations between primary care prescribing and deprivation in general medical prescribing. A recent study by Mooney et al ²⁸ found drug categories most strongly correlated with deprivation included analgesics. Conversely, both hormone replacement therapy and combined oral contraception were prescribed more in affluent areas. Ophthalmic drugs did not feature in reporting of stronger associations between either deprivation or affluency, although interestingly Latanoprost was weakly positively correlated with deprivation. Overall, Mooney et al found SES correlated with higher rates of prescriptions for a large number of drugs, with only a few drugs being correlated with affluency. For glaucoma and our study, it is pertinent to note that in contrast to conditions noted in Mooney’s study, patients with glaucoma typically remain asymptomatic, potentially even where disease is advanced but asymmetric, and care-seeking behaviour before presentation may influence prescribing in a scenario different to the management of chronic pain, for example.

In relation to the downward trend in prescribing during COVID-19, contemporaneously reduced activity in secondary care resulted in backlogs, arguably a scenario influencing patients’ behaviours around continuation with glaucoma medications, owing to follow-ups being cancelled or delayed, an explanation supported by a study assessing the impact of COVID-19 on patient-reported outcomes,²⁹ showing care perceived as being less well organised. Uncertainty among patients may have resulted in adherence failures. It is also possible that, from 2020 onwards, growth in PGA prescriptions slowed because first-line treatment was changing towards SLT. Although the timeline for our analyses pre-dated the updated guideline in January 2022 recommending SLT as first-line treatment, it is likely some ophthalmic units in England were already offering SLT following publication of high-quality evidence of effectiveness.³⁰

Implications for research and practice

The present analysis adds to earlier use of large datasets in glaucoma, for example, Saunders et al ³¹ showing the likelihood of patients suffering visual impairment in their lifetimes being linked to visual field loss at presentation, and Kelly et al ³² observing the conversion rate of OHT to glaucoma in a retrospective examination of over 45 000 electronic glaucoma records. Our time-series analysis has shown a 1-year delay in the uptake of PGAs in more deprived areas versus national data. While lower prescribing in deprived areas may be explained by a number of factors, observing quintile two has the highest rate of PGA prescribing is perhaps counterintuitive. It is uncertain if these findings reflect prescribing being unexpectedly high in quintile two, or whether prescribing is lower than expected in quintile one. We have no evidence to suggest prevalence differs between groups, and since quintile one represents affluent areas, access to sight testing should not present a referral barrier. Arguably patients in quintiles one and two may be better informed regarding health and may seek alternative management, such as SLT or minimally invasive glaucoma procedures. Furthermore, some in quintile one may obtain private prescriptions, data which would not be captured in this study, although it would be surprising if this factor accounted for the difference of 24 prescriptions per 1000 people between quintile two and quintile one in 2021–2022.

Some evidence suggests GP practices tend to be in more affluent areas, while pharmacy achieves better levels of activity in deprived areas.³³ Data for optometry in England is more aligned to the GP practice trend, with double the number of optometrists in the most affluent versus most deprived quintiles.³⁴ In Scotland, where there is a different General Ophthalmic Services contract, distribution of optometry practices is relatively balanced across SES, with Legge et al ³⁵ proposing differences in eye examination uptake across social strata is beyond service availability alone. The ophthalmology workstream, Getting it Right First Time,³⁶ recommends an optimal glaucoma care pathway; however, inequitable distribution of eyecare has potential consequences for implementation of this approach. Integrated care boards and local authorities must ensure eyecare services are in all areas. Lower prescribing of glaucoma medications in areas of greater deprivation is an unwarranted inequality. While further studies may help establish reasons for this variation, helping development of strategies to reduce inequality, GPs working in more deprived areas can play a role in promoting the uptake of sight testing for their patients at greatest risk of glaucoma.