Impact of COVID-19 pandemic on retinopathy of prematurity services in low resource settings

Abstract

Background and objectives

Retinopathy of prematurity is an increasingly important cause of blindness in children in low- and middle-income countries. Timely screening and treatment greatly reduce blindness. This study assessed the impact of the COVID-19 pandemic on ROP services in low- and middle-income countries.

Methods

An online survey with closed and open-ended questions in Microsoft Teams was sent to 304 ophthalmologists providing screening, treatment and/or vitreoretinal services (January 15 to March 12 2021). Categories were used to report impact on services. 28 participants were purposively selected for interview.

Results

184/201 completed forms from 32 countries were received. Two thirds of participants worked in government facilities and 45.6% were paediatric ophthalmologists. A moderate to severe impact was reported by 13% for screening, 9% for treatment and 16% for vitreoretinal surgery with some variation between government, private and not-for-profit sectors. 22% thought services would take a year to recover. Fifteen ophthalmologists from 12 countries were interviewed. Many reported fewer neonatal admissions, and several reported more ROP blind infants once restrictions lifted. Themes associated with services continuing included limited lockdown, autonomy and flexibility, commitment, advocacy and technology. Themes associated with a negative impact include fear and panic, closure of facilities, COVID infection or quarantining of health workers and parents, lack of transport and exacerbation of poverty.

Conclusions

The COVID pandemic had a very variable impact on the provision of ROP services. In some countries where the impact was great, more infants may have become blind from ROP. Lessons can be learnt for planning pandemic preparedness.

Introduction

More than 32,000 preterm infants became blind or visually impaired from retinopathy of prematurity (ROP) in 2010 [1]. Control requires high quality neonatal care, screening to detect vision-threatening ROP and urgent treatment when required, with follow up.

Screening can be by indirect ophthalmoscopy in the NICU by an ophthalmologist, or using widefield retinal imaging with or without remote image grading [2, 3]. Retinal images can also be taken using smartphones [4] and shared on social media. Treatment of Type 1 ROP requires urgent peripheral retinal laser photocoagulation, or intravitreal injection of anti-vascular endothelial growth factor (anti-VEGF), sometimes under general anaesthesia [5].

The COVID 19 pandemic has disrupted multiple aspects of health care through multiple mechanisms. This study was undertaken to explore the impact of the pandemic on ROP services in low, and lower- and upper- middle-income countries (LICs, LMICs and UMICs, respectively).

Methods

Mixed methods were used. After preliminary analysis of an eSurvey, 15 participants were interviewed.

eSurvey participants

Work Bank categories in January 2021 defined eligible countries [6]. Participants, all ophthalmologists, were recruited as follows: Group 1. known to one of the authors (CG); Group 2. authors of publications on ROP, and Group 3. snowballing by Group 1 contacts. All were sent an email invitation, and those responding were sent a link to the online survey (eSurvey) with at least one reminder.

eSurvey

A draught questionnaire, with closed and open questions, was pre-tested by an ophthalmologist from an UMIC and subsequently amended. The eSurvey in Microsoft Teams was sent to participants between January 15 and March 12 2021. Data were collected on where participants worked and in which sector (i.e., government, private or not-for-profit); their subspecialty, the number of NICUs where they screen and the timing and duration of lockdown in 2021. Seven categories were used to describe the impact of the pandemic on screening, treatment and vitreoretinal (VR) services (Table 1). Participants were also asked whether they had observed any changes in neonatal care and to indicate their willingness to be interviewed.

Table 1.

Categories to assess the impact of the COVID pandemic on ROP services.

1. No impact

2. No impact, but more difficult/takes longer

3. Mild/moderate impact, such as services are not as regular or frequent as before

4. Serious impact: services largely stopped for personal/financial/logistical reasons

5. Serious impact: services largely stopped as I was/am not permitted to work as before

6. Have stopped providing services, for personal/financial/ logistical reasons

7. Have stopped providing services, as I was/am not permitted to work as before

8. I did not screen/treat/provide VR services for ROP before the pandemic

Data analysis

eSurvey data were imported into Excel to check completeness. Names were used to identify duplicates but de-identified data were analysed. The data are presented as frequencies; proportions were compared using Chi square tests (Social Science Statistics, https://www.socstatistics.com).

Online interviews

Interviews were undertaken to understand factors which influenced the impact of the COVID pandemic on services. Data from the eSurvey were used to purposively select participants using (a) the impact of the pandemic (no impact through to severe impact), and (b) to ensure a range of regions, countries and health sectors. Twenty eight were invited by email, 20 of whom agreed. Fifteen were interviewed after sending reminders.

The interviews (August 2021), which lasted 20–40 min, were undertaken by one of the authors (CG) using Zoom and were audio-recorded and saved using eSurvey numbers. Participants were asked to describe their usual practices before the pandemic, and to expand on their eSurvey responses. Audio recordings were initially transcribed using speech-to-text software (otter.ai, Los Altos, California) and names or locations were redacted. Transcripts were checked against the audio recording by both authors.

Data analysis

Emerging themes were identified and manually coded by CG using inductive content analysis after immersion in the text. As interviewees were purposively selected according to whether ROP services had or had not been disrupted by the pandemic, subthemes were identified under these two main themes. Further themes and subthemes emerged from the data which were agreed by both authors after discussion. Comments in the online survey were extracted verbatim and analysed to identify new themes.

Ethics

Ethical approval was obtained from the London School of Hygiene & Tropical Medicine who waived informed consent. In the initial email invitation potential participants were asked to ascertain whether local ethical approval was required. All audio-recordings were deleted after the transcripts were finalised. The CHERRIES checklist was used in reporting [7].

Results

Quantitative findings

eSurvey participants

201/305 questionnaires were completed; 104 did not respond; 17 were excluded (nine from high-income countries; seven duplicates; one incomplete data) leaving 184 for analysis: 30 (16.4%) in Group 1, 8 (4.3%) in Group 2, and 146 (79.3%) in Group 3. Participants lived in 32 countries (Supplementary File); 103 (56%) were UMICs, 77 (42%) were LMICs and four (2%) were LICs. Most worked in government hospitals (67%), with fewer in private (24%) or not-for-profit hospitals (9%); 42% participants worked in government facilities in UMICs. Non-responders were more likely to live in UMICs than responders (66% vs 56%, respectively)(Supplementary File).

Most participants were paediatric ophthalmologists or VR surgeons (Table 2) who provided services in 929 NICUs (range 1–127). Before the pandemic, most participants screened and treated ROP, with variable provision of VR surgery.

Table 2.

Services for retinopathy of prematurity provided by eSurvey participants.

Characteristics		N	%
Subspeciality	Paediatric ophthalmologist	84	45.6%
	Vitreoretinal surgeon	78	42.4%
	General ophthalmologist	22	12.0%
Services provided for ROP	Screening only	21	11.4%
	Screening and treatment	159	86.4%
	Treatment/surgery only	4	2.2%
Provision of vitreoretinal services	Referred to another hospital	80	43.5%
	Performed by responding ophthalmologist	48	26.1%
	Performed by colleague in same hospital	28	15.2%
Provision of vitreoretinal services	Not available	24	13.0%
	Mix of the abovea	2	1.1%
	No data	2	1.1%
Total		184	100%

^aWhere multiple units were involved, different approaches were used.

Bold text denotes totals, as indicated in the row titles.

Duration of lockdown

In most countries lockdown started towards the end of March 2020, with further lockdowns thereafter in some. Lockdown lasted from 0 to 9 (mean 2.6) months with variation between economic regions (LICs: mean 1.3 months; LMICs: mean 5.8 months; UMICs mean 3.7 months).

The pandemic had no or only minimal impact (categories 1 and 2) on ROP screening (50%), treatment (61%) and VR services (65%)(Table 3). Differences between health sectors did not reach statistical significance (Fig. 1A). However, differences were more pronounced between World Bank regions (Fig. 1B). No impact on neonatal care was reported by 47% participants (some changes, 39%; major changes, 10%). Differences between health sectors were not statistically significant. Approximately equal proportions considered recovery of ROP services would take a few weeks (18%) or up to a year (22%).

Table 3.

Impact of the COVID pandemic of services for ROP in low- and middle-income countries amongst eSurvey participants and the subset who were interviewed.

	sSurvey participants (184)						Subset interviewed [15]
Category	Screening		Treatment		VR surgery
None/minimal [1, 2]	92	50%	112	61%	120	65%	5	33%	7	47%	9	60%
Mild/moderate [3]	66	36%	47	26%	30	16%	4	27%	3	20%	3	20%
Severe [4, 5]	22	12%	15	8%	30	16%	3	20%	4	27%	-	-
Stopped [6, 7]	2	1%	1	1%	0	0%	3	20%	0	0%	1	7%
No data [8, 9]	2	1%	9	5%	4	2%	0	0%	1	6%	2	13%
Total	184	100%	184	100%	184	100%	15	100	15	100	15	100

Bold text denotes totals, as indicated in the row titles.

Fig. 1. Impact to COVID 19 pandemic on services for ROP.

(A) By health sector (B) By World Bank Economic Region.

Qualitative findings

Participants interviewed

In the eSurvey 28 participants declined interview and non-English speakers were excluded; 28 of the remaining 154 participants were purposively selected. Twenty agreed to be interviewed and were sent a date; five did not respond. Fifteen participants from 12 countries were interviewed (response rate 54%). See Supplementary File for the characteristics of participants interviewed.

Participants provided ROP services in 184 NICUs (range 1–127, median 2). The first lockdown ranged from 0 to 6 (mean 3.6, median 3) months and the restrictions imposed also varied. Some were involved in longstanding ROP programmes whereas others had recently started services. In the eSurvey, five (33%) participants reported that the pandemic had no impact ROP screening whereas three (20%) reported that screening stopped (Table 3). The impact was less marked for treating ROP and on VR services.

Findings are presented under the following headings: 1. impact of the pandemic overall; factors which 2. enabled services to continue or 3. had a negative impact on services 4. resumption of services, and 5. other comments.

Impact overall

Number screened

Several participants reported a fall in the number of babies screened, particularly after discharge from the NICU. The first quote is from an ophthalmologist who manages a large telemedicine screening programme.

“We averaged 2500–3000 screenings a month pre COVID. In April 2020 this was down to 650” (48 eSurvey (eS) quote/LMIC)

Number treated

The impact on the number of infants treated for acute ROP or who had VR surgery varied:

“Any usual year….we treat approximately 50 babies….Last year….it was 25 or 30.” (79/LMIC)

“The proportion of babies requiring treatment was the same despite the pandemic.” (48/LMIC)

“There was a higher incidence of ROP and severe ROP in my hospital…..compared to previous years” (4eS/UMIC)

In one setting, an increase in referrals from ophthalmologists who stopped treating during the pandemic led to a marked increase in the number of infants treated for two to three months.

More ROP blind babies

Several ophthalmologist reported that they saw more ROP blind infants once travel restrictions lifted as parents could travel.

“June to August we were flooded by babies with advanced disease that were not screened or treated on time in neighbouring states….” (48eS/LMIC)

“….screening was not very effective in 2020. As a result, many patients have gone blind due to lack of proper screening” (41eS/LMIC)

Factors which enabled ROP services to continue

Not a strict lockdown

All but one interviewee described strict lockdowns. In this country physicians were permitted to travel freely after showing their institutional identity.

Autonomy and flexibility

Ophthalmologists providing ROP services in their institution’s NICU could continue as travel was not required. Participants working in the private sector reported several advantages, as the NICUs tended to be smaller with fewer visitors, which reduced the risk and fear of COVID.

Another private hospital created a separate facility for COVID patients, which reduced infection amongst staff in the main hospital. An ophthalmologist from a not-for-profit hospital, who screens in government and private NICUs, commented that the impact was much greater in government than in private facilities, as the latter were not included in government COVID control strategies. Families of private patients were also better educated and wealthier and could adhere to advice after discharge from the NICU.

“Private [patients] were from more educated and well to do families - they can come to our hospital with their own transportation. The government sector of poor patients were late to reach us compared to private babies.” (79/LMIC)

Commitment

Several participants showed great commitment. For example, one drove long distances to provide services when domestic flights stopped, while others covered for colleagues who were sick, vulnerable or not willing to screen.

Advocacy

Several continued providing services because ROP was classified as an emergency. In one country, the ophthalmologists successfully advocated that ROP be classified an emergency, which allowed services to continue.

Technology

Several participants used telemedicine or took images using Smartphones and were often part of a network who communicated using social media. This enabled them to share images and plan how to cover colleagues who were unable to provide services. In one country, all 99 screening ophthalmologists were connected in a network, and could communicate with the three ophthalmologists who could treat ROP.

Factors which had a negative impact on ROP services

A range of inter-connected factors led to negative impacts on ROP services (Fig. 2).

Fig. 2.

Interconnected pathways in which the COVID pandemic negatively impacted services for ROP.

Fear and panic

Several participants described great fear and panic amongst parents, NICU staff and ophthalmologists when the pandemic started, as so much was unknown.

“When COVID came, there was a lot of panic and confusion. Everyone [neonatal unit staff] was panicking [saying] ‘Okay, doctor, okay. But people are dying from COVID. Do you think that [ROP screening] is important while we have this problem?’ ” (201/LMIC)

“There was tremendous panic, even amongst healthcare workers. We were not clear about the mode of transmission. Our team was feeling as if we are going onto a battlefield, because we didn’t know what kind of risks we were carrying” (7/LMIC)

"The factor which [caused the] most panic was… in babies, you can’t put a mask on their face, and they’re crying all the time. And there was a lot of discussion about this droplet transmission and how you’re going to prevent it …when you’re screening or treating [ROP”] (7/LMIC)

One participant described the protective clothing he wore during early lockdown when he visited another NICU to treat an infant using laser.

“..it was like a war zone…I had to wear something.. like an astronaut…No one knew how the disease spread and everyone believed that if you were not in that suit you would either get or give the disease.” (48/LMIC)

One infant with advanced ROP went blind because the anaesthetic staff refused to give a general anaesthetic.

A reduction in NICU admissions during lockdown, and often for a prolonged period thereafter, was reported by most participants as there was fear of attending hospitals.

“Institutional deliveries were markedly decreased during the first 3–4 months of COVID.” (201/LMIC)

“In the last week of March 2020 there was mandatory leave and widespread panic which forced doctors and patients to stay at home. The number of neonates in the NICUs were reduced drastically in all the NICUs I visited initially” (150eS/LMIC)

Closure of facilities

Early in the pandemic some government hospitals or their NICUs closed, while others were converted to COVID hospitals. In some instances, staff were transferred from NICUs to care for patients with COVID. These factors had a major impact on all aspects of ROP services.

“When the lockdown came the university closed… this whole [ROP] service stopped” (11/LMIC)

“Of the 17 services that I attended before the pandemic, three ended the activities of the NICU to accommodate COVID beds.” (179eS/UMIC)

“We used to refer babies with Stage 4 and 5 (ROP) [for VR surgery]. However, the ….institute …was converted to a designated COVID hospital….” (102eS/LMIC)

Staff infection

Sickness amongst hospital staff also had a major impact on services in some settings.

“The other [maternity] hospital…about 80% of [staff] caught COVID within a week. It was a Ministry of Health emergency - the whole hospital was closed down - all the patients were discharged.” (11/LMIC)

“Screening stopped in one unit as several staff became infected and this led to closure. The kangaroo mother’s room was closed, reducing numbers of babies screened drastically.” (11eS/LMIC)

COVID infection amongst NICU staff meant that remaining staff were overworked and could not assist ROP screening.

“…nurses just forgot to dilate the babies [pupils] because they were overwhelmed with so many babies..” (7/LMIC)

Ophthalmologists also developed COVID. In one eye department, the screening ophthalmologist was the head of department. COVID infection amongst the residents because of inadequate PPE meant his focus was on keeping the eye department running.

Quarantine and PCR testing

Health care professionals and parents sometimes had to quarantine after contact with an infected individual.

“One infant’s father’s was detected with COVID the day before [VR] surgery and the whole family was quarantined….. for 17 days…By the time they came back, it was inoperable.” (7 LMIC)

Several participants reported that later in the pandemic PCR testing was required, particularly before ROP treatment or VR surgery, which also caused delays.

Lack of transport

Lack of public transport and/or travel restrictions during lockdown had a major impact on NICU staff, ophthalmologists, and parents of infants who needed screening or treatment after discharge. Several ophthalmologists were given passes/permission to travel within their city or district but could not travel further afield.

“Transport completely stopped. And we needed government passes to move….but patients were unable to reach the centre for screening.” (79/LMIC)

“…inter-district travel was completely banned. Twice our vehicle (for the screening team) was sent back by the police at the district border so there was nothing we could do.” (7/LMIC)

Several said that the cost of private transport was prohibitive for poor families.

“…follow up was a very big challenge….Many of them could not afford [a taxi].” (34/LIC)

Some infants with Type 1 ROP could not move from one district to another for treatment.

“Babies referred from outside my city sometimes took longer to get to me….and some progressed to higher stages.” (20eS/LMIC)

One participant gave a graphic account of the impact of the inter-district travel ban on access to treatment. In this setting, some ophthalmologists refer infants needing treatment to another district. However, during the pandemic when local treatment was not available, some babies were treated by inexperienced ophthalmologists.

“There were a lot of incomplete treatments, late treatments who were coming to us with bad-looking ROP. In some we could just salvage one eye after treatment. It was like dealing with victims of war.” (48/LMIC)

Several African participants had only recently started ROP services. Because they did not have lasers, they used anti-VEGF agents which often had to be sourced from overseas. Restrictions on air travel made this even more difficult.

Poorer quality neonatal care

Shortages of NICU staff, for whatever reason, led to worse nurse to baby ratios. Several participants observed poorer quality of care, particularly in oxygen administration.

Poverty exacerbated

During the pandemic people not formally employed or in non-essential jobs were particularly vulnerable to loss of income. This, combined with the need to pay for private transport, impacted parent’s ability to access treatment.

“Quite a few parents said they had lost their job or their income had reduced, and they could not afford treatment. Even if we said ‘treatment would be free, you just come’, they just didn’t have money to come” (7/LMIC)

Change in treatment

Several ophthalmologists modified how they treated infants. Fear of COVID transmission during laser treatment meant that some used anti-VEGF agents instead, which is quicker. Some started using laser after anti-VEGF treatment as follow up after monotherapy could not be guaranteed.

Resumption of services

Several participants reported that ROP services had almost returned to normal. Others thought it would take longer, or would be difficult, as COVID vaccination was being prioritised by Ministries of Health and the backlog of ophthalmic outpatients hampered resumption of normal ROP services.

Other comments

Several ophthalmologists said that plans to expand ROP services stalled during 2020 as they could not travel to train, or the authorities were focusing on COVID. One participant commented on the advantages that telemedicine would have offered during the pandemic, as one trained individual could visit several NICUs.

Some reported positive impacts in relation to neonatal care, and in one country the screening guidelines were modified.

“[Nurses worked in] “bubbles”….We noticed we had less residents and observers and less sepsis cases. The units were quiet, no noise. Communication between the NICU team (members) was stronger.” (5/UMIC)

“We streamlined the screening [guidelines] and created a lot of awareness among people to take on ROP (services)…. There’s a lot of sharing of information now….” (48/LMIC)

Discussion

The COVID pandemic affected health in many ways globally both directly and indirectly. Routine health services were often curtailed during lockdown, including surgery for congenital conditions [8, 9]. We identified six studies on the impact on ROP services, from the USA [10], India [11–14] and China [15]. Two described the mitigation measures adopted which allowed services to continue [10, 14]. The other studies all reported that fewer infants were screened during lockdown, with a variable impact on the number treated. This may reflect publication bias, as half of the eSurvey participants in our study reported no or only minimal impact on screening or treatment. Our study is the first to report that in some countries more preterm babies were likely to have gone blind from ROP because of the pandemic, but participants did not have data to support their statements.

Fewer NICU admissions during lockdown were reported by many participants and in publications [14]. This is likely to reflect fear and panic, COVID infection among staff, NICUs being converted to COVID care and travel restrictions. This may, in part, explain why some participants reported fewer infants needing treatment. However, several high-income countries and China have reported lower preterm birth rates during lockdown [16–19], which has led to international collaborative research on interventions to prevent preterm birth [20].

In our study, the impact on screening was greater in government than in private hospitals, but differences were not statistically different, which may reflect the small sample size. Explanations given were that families accessing private neonatal care could afford to travel, neonatal staff were not transferred for COVID duties, and managers had greater autonomy to make changes. Differences between economic regions were more pronounced which may, in part, reflect the scarcity of ophthalmologists skilled in ROP, particularly in LICs in Africa [21].

Half of the participants were able to provide ROP services throughout lockdown despite PPE making this more difficult, which is testament to their commitment. Several mentioned the role of social media, which ensured continuity of care when screeners became sick, or were quarantining or vulnerable. Others thought that telemedicine would be advantageous, but there are considerable resource and infrastructure implications [22].

Several ophthalmologists changed how they treated ROP during lockdown, which was influenced by lack of general anaesthesia because anaesthetists were either caring for COVID patients or they feared droplet transmission. Another factor was anxiety that infants treated with anti-VEGF would not return regularly for follow up. To address this challenge, some started routine laser a few weeks after anti-VEGF treatment.

Several African ophthalmologists said they routinely use anti-VEGF agents as lasers are not available [21], but accessing them was often difficult during the pandemic. There is an urgent need to ensure that all treatment modalities for ROP are accessible and affordable. The impact on VR services is more difficult to interpret, as several participants referred infants for surgery before the pandemic and continued to do so during the pandemic.

In LMIC and LIC parents of preterm infants are often poorly educated and are poor, and compliance with outpatient screening and treatment is challenging even under normal circumstances [23, 24]. Loss of income during the pandemic would have made compliance even more difficult, particularly for those in the informal labour market. The economic impact of the COVID pandemic on households and the ensuing recession [25] are likely to continue to affect access to health care, including for ROP services. Ensuring that ROP services are covered by health insurance would alleviate the economic burden on parents.

There are several limitations in our study. Response rates to the eSurvey and interviews were low which may have introduced selection bias. Non-responders to the interview were not replaced, as no new themes emerged during the last few interviews. Almost half of those interviewed were known to the interviewer, which may have introduced bias, and including neonatal staff and parents would have given a more complete picture. Another limitation is that most ophthalmologists could not provide data on the number of infants screened or treated before and during the pandemic.

Conclusion

This study highlighted the limitations of ROP services in some locations, such as lack of lasers, and the limited number of ophthalmologists with skills to treat ROP, which accentuated the impact of the pandemic. The study also highlighted lack of pandemic preparedness, and how mitigation strategies were developed as the pandemic evolved. Ideally ROP screening and treatment should be included as an essential service in pandemic preparedness plans, as in India [26]. This might include adapting national ROP guidelines, advocating that infants requiring treatment be permitted to travel and developing networks of ROP service providers to maintain continuity of care.

Summary

What was known before

• Few studies have addressed the impact of the COVID pandemic on services for ROP screening and treatment, and on the provision of vitreoretinal surgery.

• Most report a fall in the number of preterm infants screened.

What this study adds

• In many centres ROP services continued despite being more difficult.

• Participants reported screening fewer preterm infants and some reported more ROP blind infants.

• Interconnected factors which negatively impacted services are described, and implications for pandemic preparedness are outlined.

Author contributions

CEG conceptualised and designed the study, collected and analysed the data and wrote the first draught of the manuscript and takes responsibility for the intellectual content. JT assisted in data management and analysis and reviewed and revised the manuscript.

Data availability

Data are available from the corresponding author upon request. Deidentified individual participant data will not be made available.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41433-023-02784-6.