Equity reporting in systematic reviews and meta-analysis for geographic atrophy: a PROGRESS-Plus assessment

In developed countries such as the United States, Australia, Japan, and much of Western Europe, age-related macular degeneration (ARMD) is the leading cause of registered legal blindness in individuals older than 65 [1]. In 2019, it was estimated by the Centers for Disease Control and Prevention (CDC) that 19.8 million Americans are living with the disease [2]. As ARMD progresses, there is a greater risk of developing either the neovascular form or geographic atrophy (GA), both of which are vision-threatening late sequelae. GA, in particular, is an irreversible and centrally blinding manifestation characterized by sharply demarcated hypopigmented lesions that usually appear first in the perifoveal macula [3]. Over time, these lesions will likely expand into areas of drusen, or whitish-yellow deposits within the retina, and eventually into the fovea [4]. Reticular pseudodrusen, a particular subtype of drusen located above the retinal pigment epithelium (RPE), has been more closely associated with GA development and multilobular atrophy expansion [3], [4], [5], [6]. Visualization of atrophic lesions can be done by a fundus examination or other imaging modalities. On optical coherence tomography (OCT), GA is identified by RPE atrophy and light hypertransmission through the retinal layers, whereas on fundus autofluorescence (FAF), these lesions appear as hypofluorescent patches due to their lack of lipofuscin [3], 6]. The size and growth of GA can also be monitored with regular examinations and imaging, FAF being particularly recommended for this purpose in the literature [5]. The rates of GA expansion are highly correlated with the age of disease onset, as those diagnosed at a younger age are likely to have a poorer prognosis [5]. As atrophy progresses into the fovea, visual acuity is significantly decreased due to the development of irreversible central scotomas [7].

Several factors play a role in the advancing size and progression of GA. Increasing age and a family history of the disease magnify the risk of developing this sequela [7], [8], [9]. A predilection based on gender is still debated in the literature. Although some articles suggest that this may be a male-predominant disease, other studies have found no statistical significance between males and females [8], 9]. Among different ethnic populations, it has been shown that those of European or Native American ancestry are more likely to develop GA, whereas the Asian population is less likely to do so [8], 10].

Modifiable risk factors for GA include smoking and hypercholesterolemia [11], 12]. Smoking has proven to increase the risk of disease onset as well as hasten lesion growth once present, whereas quitting has demonstrated the ability to decrease GA development [7], 13]. While research has shown that eating low-glycemic-index foods such as those found in the Mediterranean diet can protect against GA enlargement [14], 15], individuals experiencing food insecurity – defined by the United States Department of Agriculture as “a household-level economic and social condition of limited or uncertain access to adequate food” – are less likely to be able to obtain these foods, and they are not able to afford antioxidant vitamins that have shown to slow moderate ARMD progression before the disease progresses to later stages such as GA [4], 16]. Although a correlation between physical activity and progression to late-stage ARMD has been debated in the literature, overall incidence of the disease has proven to be lower among individuals with higher physical activity levels [17], 18]. Other conditions, such as diabetes or cardiovascular disease, are associated with ARMD because many risk factors for all three diagnoses are the same [19]. Altogether, the current research hypothesis is that all of these factors alter complement and mitochondrial function, increasing or mitigating oxidative stress and contributing to disease pathogenesis or prevention [20].

Systematic reviews and meta-analyses (SRMAs) that address GA and disease epidemiology are of utmost importance to physicians when developing treatment plans. These documents summarize the most recent literature and are frequently used as a starting point for disease management [21]. With this in mind, creating SRMAs that apply to all populations is crucial because they can promote equitable care by providing more information to the doctors who read them. Equitable care, in turn, may lead to decreased disease burden. These facts have not been overlooked by the American Academy of Ophthalmology (AAO), which predicted that creating an environment that reduces disparities may significantly reduce the projected growth of visual impairments and blindness in the Unite States by the year 2050 [22], 23]. To take this one step further, the AAO has created a task force that produces commentaries and publications that highlight how social determinants such as education, healthcare accessibility, health literacy, and even ZIP code can have a significant impact on the care given to patients and their visual outcomes [22]. As these initiatives are underway, it remains unclear whether the literature utilized to manage GA reflects this hard work directed toward increasing equity.

A framework commonly utilized to assess equitable reporting in SRMAs is called PROGRESS. It was developed by Evans and Brown in 2003 and is a method by which to analyze literature and draw conclusions regarding its holistic reporting or lack thereof [24], 25]. The acronym PROGRESS stands for Place of residence, Race/ethnicity/culture/language, Occupation, Gender/sex, Religion, Education, Socioeconomic status, and Social capital [24]. The original framework was modified to PROGRESS-Plus – allowing additional categories including personal characteristics associated with discrimination (e.g., age, disability), features of relationships (e.g., smoking parents, excluded from school, smoking in the home), and time-dependent relationships (e.g., leaving the hospital, respite care, other instances where a person may be temporarily at a disadvantage), etc. [25], 26] For this study, we employed the PROGRESS-Plus method to analyze SRMAs concerning GA, because it is unknown whether the current literature comprehensively addresses all social determinants of health. We hypothesize that a majority of SRMAs pertaining to GA do not adequately report equity-related items found in this framework. Given the detrimental effects of GA to patients’ vision and the many new treatment advancements being made specifically for it, it is important to have an equitable research foundation for further advancement. By analyzing this, we hope to provide more insight to those researching and treating GA, as well as to potentially reduce the predicted future disease burden.

Results

Our initial search string yielded 176 articles, of which 119 were excluded because they were duplicates or did not meet the eligibility criteria. Figure 1 is a PRISMA flow diagram that shows the breakdown of the screening process with the reasons for exclusion.

Figure 1:

A PRISMA flow diagram showing the breakdown of the screening process with reasons for exclusion.

All of the data from the search string and analysis have been uploaded to Open Science Framework and can be viewed with the following link: https://osf.io/hbafm/?view_only=928da914008d4f74aaa5e1a9cb2e53d2.

Equity inclusion within systematic reviews

Of the 57 SRMAs included in our sample, 31 (55.4 %) included at least one element of the PROGRESS framework with age, sex, and gender most widely distributed – leaving 26 (45.6 %) articles that did not include any PROGRESS-Plus items (Table 1). An additional 26 articles contained less than five of the extracted elements but named at least one. Only 4 (7.0 %) articles contained 5–9 items, and only one (1.8 %) contained 10 items (Table 1). The most commonly cited constituents were age, which was listed in 43.9 % of analyzed SRMAs, gender/sex (29.8 %), and race/ethnicity (28.1 %) (Table 2). Smoking status was listed in 7 (12.3 %) SRMAs, and nutritional status/food security was listed in 2 (3.5 %) (Table 2). Religion and the presence of others smoking in the household were not cited in any article, and occupation, socioeconomic status (SES), social capital, education, disability, and a sedentary lifestyle were only listed in one SRMA (Table 2).

Table 1:

The frequency of PROGRESS-Plus items included within systematic reviews (n=57).

PROGRESS-Plus item inclusion	No., %
0	26 (45.61)
1	11 (19.3)
2	3 (5.26)
3	5 (8.77)
4	7 (12.28)
5	2 (3.51)
6	1 (1.75)
7	1 (1.75)
8	0 (0)
9	0 (0)
10	1 (1.75)

Table 2:

Systematic reviews’ inclusion of PROGRESS-Plus items (n=57).

PROGRESS-Plus item inclusion	No., %
Place	14 (24.56)
Race/ethnicity	16 (28.07)
Occupation	1 (1.75)
Gender/sex	17 (29.82)
Religion	0 (0)
Education	1 (1.75)
Socioeconomic status, SES	1 (1.75)
Social capital	1 (1.75)
Age	25 (43.86)
Disability	1 (1.75)
Smoking status	7 (12.28)
Others smoking in HH	0 (0)
Food security/nutritional status	2 (3.51)
Sedentary lifestyle	1 (1.75)
Other comorbidities	6 (10.53)

1. HH, health history.

Additional study characteristics

Among the study characteristics that were extracted alongside PROGRESS-Plus items, the location of the study was the only one found to be statistically significant (p=0.03). Of the 31 SRMAs that contained at least one PROGRESS-Plus item, 26 of them were conducted outside of the United States (Table 3). Over two-thirds (68.8 %) of the SRMAs conducted within the United States did not contain any PROGRESS-Plus elements, compared to 36.6 % in non-US countries. First-author affiliations, study funding, presence of patient-reported outcome measures, and specifically mentioning PRISMA-equity or PROGRESS within the articles were not found to be statistically significant (Table 3).

Table 3:

Associations between inclusion of any PROGRESS-Plus items and study characteristics.

	Does not include any PROGRESS-Plus items No., %	Include PROGRESS-Plus items No., %	, χ 2, P
Total	26 (45.61)	31 (54.39)	–
Location of study
Non-United States	15 (36.59)	26 (63.41)	4.80, 0.028
United States	11 (68.75)	5 (31.25)
Affiliation of first author
Government or government agency	0 (0)	1 (100)	6.05, 0.11
Hospital or medical institution	4 (23.53)	13 (76.47)
Industry	1 (50)	1 (50)
University	21 (56.76)	16 (43.24)
Funding for study
Government or government agency	4 (25)	12 (75)	4.92, 0.30
Hospital, medical institution	3 (37.5)	5 (62.5)
Industry	6 (54.55)	5 (45.45)
No funding	3 (60)	2 (40)
Not available	10 (58.82)	7 (41.18)
Mentions PRISMA-Equity or PROGRESS framework within article
No	25 (48.08)	27 (51.92)	1.45, 0.23
Yes	1 (20)	4 (80)
Did the SR include any patient-reported outcomes as assessment measures?
No	23 (46)	27 (54)	0.02, 0.88
Yes	3 (42.86)	4 (57.14)

1. PRISMA, preferred reporting items for systematic reviews and meta-analyses; SR, systematic review.