Time to treatment of pediatric retinal detachments: A United States Claims-Based Analysis

Abstract

Purpose:

To characterize delays in diagnosis and treatment of retinal detachments (RD) in a pediatric population.

Design:

retrospective cohort study using insurance claims data

Subjects:

pediatric patients with retinal detachment who underwent repair in the outpatient setting

Methods:

A retrospective analysis of commercially insured patients from a national cohort (IBM MarketScan Research Databases), age 18 years and younger with an incident diagnosis of retinal detachment between 2007–2016. Patients with preceding eye-related visits, time to diagnosis, and time to repair were calculated and compared between patients with pre-existing ocular diagnosis and those without.

Main outcome measures:

time from diagnosis to specialist consultation, time from diagnosis to repair, time from specialist consultation to repair, number of preceding visits, presence of previous eye-related diagnosis

Results:

Our sample consisted of 826 patients, the majority of whom were diagnosed with rhegmatogenous RD (77%). Only 40% of patients had at least one preceding eye-related visit and 33% had at least two visits prior to RD diagnosis with a median time from the last eye-related visit of 32 days (4–197) and median time from second to last visit of 118 days (24–437). Median time from RD diagnosis to repair was 2 days (0–9). The 323 (37.9%) patients with pre-existing ocular diagnoses more frequently had at least one (44% vs. 37%, p=0.079) or two preceding eye-related visits (40% vs. 29%, p=0.002) relative to those without and also had a shorter time to RD diagnosis (median 14.5 days vs. 44.5 days, p=0.011) and repair (1 day vs. 3 days, p=0.003).

Conclusions:

Retinal detachment is an important cause of morbidity in children. This work highlights how pediatric patients without previous ocular diagnoses and/or visits with eye care professional may have a delayed diagnosis and repair of their retinal detachment.

Introduction

Retinal detachment (RD) is a sight threatening condition that may be characterized as rhegmatogenous, tractional, exudative, or some combination therein.¹ Rhegmatogenous RD (RRD) is the most common subtype with a variable incidence among the published literature (up to 0.7 in 100,000 children).^1–5 Children with RD may have pre-existing systemic diseases (e.g Stickler syndrome) and ocular conditions (e.g. myopia) which could be picked up on screening exams. However, RRDs in children are often secondary to trauma^{1, 3, 4}, which cannot be prevented by any screening program. Pediatric patients with RD will nevertheless present late due to the slower progression of vision loss relative to adults and the inability of young children to clearly report their symptoms.^{1, 5}Therefore, they frequently face the clinical features of longstanding RD, including worse visual acuity (VA), macular involvement, and proliferative vitreoretinopathy (PVR).^{3, 4} Consequently, compared to adults, pediatric patients require different surgical approaches, and have worse functional outcomes thereafter, with a higher lifelong risk of recurrent RD, glaucoma and cataract.¹

Visual impairment secondary to pediatric RD results in poor vision-related quality-of-life and amblyopia, which can negatively impact educational attainment, job choices, psychosocial development, self-perception, and family well-being.^6–9 Early diagnosis and treatment of pediatric RD is essential to decrease the burden of this condition. However, the literature regarding delays in presentation and repair of pediatric RD is limited to case reports and small case series.^{10, 11} Larger population-based cohort data is currently lacking on this subject, and conclusion are therefore hard to draw from these small studies.

The aim of this study is to characterize lead-time in diagnosis and treatment of RD in a pediatric population in the United States using a national claims-based database. We additionally identified patients with pre-existing ocular diagnoses to distinguish differences in time to RD diagnosis and repair among patients with established care relative to those without.

Methods

We conducted a retrospective, cross-sectional analysis of patients from a national database of commercially insured enrollees who received a diagnosis of an incident RD from 2007 to 2016 in the United States. This study was exempted by the Stanford University Institutional Review Board and was conducted in adherence to the tenets of the Declaration of Helsinki.

Data source and patient population

We selected our cohort from the IBM MarketScan Commercial Claims and Encounters database, which are a family of de-identified patient-level medical and drug insurance claims datasets. These databases include a sample of more than 150 million Americans with employer-sponsored commercial health plans and contain records of hospital stays, outpatient services, diagnosis codes, procedure codes, demographic information and payment information. In this study, we included patients aged 18 years and younger with continuous enrollment during the study period, a medical claim from 2007–2016 (period available in MarketScan during time of data analysis) associated with an incident diagnosis of RD who underwent repair with cryotherapy, laser photocoagulation, pneumatic retinopexy (PR), scleral buckle (SB), or a pars plana vitrectomy (PPV) with or without a scleral buckle. We initially restricted our sample to patients with at least 2 distinct RD diagnosis claims, at least 182 days (or 6 months) of enrollment preceding their first diagnosis of RD to minimize the inclusion of patients with a pre-existing RD, and at least 182 days (or 6 months) of coverage after the diagnosis. “Pre-existing ocular diagnoses” were identified as those conditions increasing risk of retinal detachments. These diagnoses included myopia, trauma, aphakia/pseudophakia, YAG laser for posterior capsular opacification, lattice degeneration, retinoschisis, Marfan syndrome, Ehlers-Danlos syndrome, Familial Exudative Vitreoretinopathy (FEVR), congenital corneal malformation, congenital vitreous abnormality, retinopathy of prematurity, and other congenital malformation.

All diagnoses were identified based on the International Classification of Disease, Ninth and Tenth Revision, Clinical Modification (ICD-9-CM and ICD-10-CM) and all procedures based on Current Procedural Terminology (CPT) codes. Our cohort was representative of patients diagnosed in the outpatient setting and we excluded any patients with an RD repair procedure that preceded their first diagnosis of RD or if they did not receive an RD treatment within 6 months of their incident diagnosis (n=456, 35.6%).

Eye-related visit definition

Physician identifiers were drawn from MarketScan’s provider identification variable using the encounter associated with a given patient’s eye-related visits preceding their incident RD diagnosis. We estimated provider specialty and subspecialty using procedure and diagnosis codes by using a method previously employed for the same database.¹²

Statistical analysis

Cohort characteristics were described with medians and interquartile ranges (IQRs) for continuous variables and proportions for categorical variables. We performed descriptive analyses to characterize time from eye-related provider visit to diagnosis and diagnosis to treatment, as well as demographics, pre-existing ocular conditions, and treatment. Categorical variables were compared using Pearson’s chi-square or Fisher’s exact tests and continuous variables were compared using a Wilcoxon-Mann-Whitney test. In all analyses, P-value <0.05 was considered statistically significant. All statistical analyses were conducted using R 3.6.1 (R foundation for Statistical Computing, Vienna, Austria).

Results

Study sample characteristics

Our sample consisted of 826 patients under the age of 18 with incident RD. Nearly two thirds of patients (62.0%) were male with a median age of 15 years old, and 59.8% were born at term (gestational age of 37 weeks or later). Forty percent (n = 332) of patients were preterm, but only 2.4% were born at a gestational age of less than or equal to 30 weeks. The majority of patients had a diagnosis of a rhegmatogenous retinal detachment (77.0%). The most common procedure performed was PPV with or without a SB (64.4%), followed by SB (31.5%), laser barricade (11.7%), PR (3.1%), and cryotherapy (1.3%).

Ocular comorbidities and time to diagnosis and repair

The three most common ocular comorbidities were trauma (26.3%), aphakia/pseudophakia (23.5%), and lattice degeneration (21.9%) (Table 1). Among all patients, only 39.5% had at least one preceding eye-related visit prior to the RD diagnosis and 33.1% had at least two preceding eye-related visits. The median time from the last eye-related visit to RD diagnosis was 32 days (IQR 4–197) (Figure 1A) and the time from second to last eye-related visit was 118 days (IQR 24–437) (Figure 1B). Median time from RD diagnosis to repair was 2 days (IQR 0–9) (Figure 2).

Table 1: Characteristics of Patients with Retinal Detachment under the age of 18.

Baseline demographic data, diagnosis and treatment characteristics of pediatric patients presenting with RD

	Total (n=826)
	n/median	%/IQR
Sex
Male	512	62.0%
Female	314	38.0%
Age Group
0–5 years	74	9.0%
6–10 years	114	13.8%
11–18 years	638	77.2%
Type of RD
Rhegmatogenous	636	77.0%
Serous	48	5.8%
Tractional	142	17.2%
Pre-existing ocular condition
ROP	50	6.1%
High myopia	101	12.2%
Trauma	217	26.3%
Aphakia/Pseudophakia	194	23.5%
YAG	62	7.5%
Lattice degeneration	181	21.9%
Retinoschisis	43	5.2%
Marfan Syndrome	16	1.9%
Ehlers-Danlos Syndrome	2	0.2%
FEVR	46	5.6%
Congenital corneal malformation	14	1.7%
Congenital vitreous abnormality	19	2.3%
Other congenital malformation	45	5.4%
Term Status
Preterm a	332	40.2%
Term	494	59.8%
Treatment
Pneumatic retinopexy	26	3.1%
Scleral buckle	260	31.5%
Pars plana vitrectomy with or without scleral buckle	532	64.4%
Laser barricade	97	11.7%
Cryotherapy	11	1.3%
Proportion with preceding eye-related visit	326	39.5%
Proportion with at least two preceding eye-related visits	273	33.1% 4-
Time from last eye-related visit to RD diagnosis, days	32	197.2 24-
Time from second to last eye-related visit to RD diagnosis, days	118	437
Time from RD diagnosis to RD repair, days	2	0–9

^aPreterm was defined as patients born at a gestational age of less than 37 weeks

^bRD, retinal detachment; IQR, interquartile range; ROP, retinopathy of prematurity; YAG, yttrium-aluminum-garnet; FEVR, familial exudative vitreoretinopathy

Figure 1.

A: Histogram of time from last eye-related visit to retinal detachment diagnosis for all patients

B: Histogram of time from second to last eye-related visit to retinal detachment diagnosis for all patients

Figure 2:

Histogram of time from retinal detachment diagnosis to repair for all patients

Differences between patients with and without pre-existing ocular conditions

Three hundred and thirteen patients (37.9%) had a previously diagnosed ocular condition (specifically conditions increasing risk of retinal detachments). There were no differences in the proportion of patients by sex, type of retinal detachment, or type of treatment relative to the 513 (62.1%) patients without a previously diagnosed condition. However, previously diagnosed patients were younger at the age of their RD diagnosis (median age 14 vs. 15, p<0.001) and were more frequently born preterm (46.6% vs. 36.3%, p=0.004) compared to those without a previous diagnosis. Importantly, more patients with a pre-existing ocular diagnosis had at least one (43.5% vs. 37.0%, p=0.079) or two preceding eye-related visits (39.6% vs. 29.0%, p=0.002) relative to those without. The median time from the last eye-related visit to RD diagnosis was 14.5 days for patients with pre-existing ocular diagnosis and 44.5 days for those without (p=0.011). Similarly, median time from second to last visit was shorter for those who had previously been diagnosed with a pre-existing ocular condition (70.5 days vs. 159 days, p=0.005). Furthermore, RDs in patients with a pre-existing condition were repaired sooner than those without (1 day vs. 3 days, p=0.003).

Discussion

In this study, we explored the time between diagnosis and treatment of retinal detachments (RD) in a pediatric population. We found that those patients with an identified pre-existing condition tended to have more frequent follow ups (time from second to last eye-related visit to RD diagnosis 70.5 versus 159 days, p=0.0005) but also significantly shorter time from the last eye-related visit to RD diagnosis (time from last eye-related visit to RD diagnosis 14.5 versus 44.5 days, P=0.011) compared to those without a pre-existing condition. Of all the patients with RD diagnoses under the age of 18 years old, 62.1% were found to have had RDs in the absence of a known diagnosis of a pre-existing condition. The significant delay in time to diagnosis of their RD found in this study implies a need for improvement in eye health screening protocols for this population. History of trauma, aphakia/pseudophakia, and lattice degeneration composing most of the pre-existing ocular conditions in this study also highlights the need to consider more frequent screening in select populations as well as potentially earlier prophylactic treatment of high risk retinal lesions like lattice in young patients.

Pediatric eye screening is important in the early diagnosis of retinal detachment. However, traditional pediatric screening does have limitations. For example, neonatal and infant eye screening is administered by neonatologists and pediatricians by means of the red reflex test which is limited in its ability to detect posterior segment pathology and has varying sensitivity and specificity intrinsic to the operator of the ophthalmoscope.^13–17 Red reflex screening may also detect significant refractive error, but requires primary care providers to be attune to more subtle changes in the red reflex, as opposed to the presence or absence of leukocoria. Universal newborn eye screening with fundus photography is one possible answer.¹⁸ Periodic vision screening in school-age children is also subject to great variation in reliability based on the age group getting screened, the person administering the test, as well as the follow-up and referral pipeline criteria.¹⁹ While guidelines exist to identify children and adolescents with vision impairments^20–22, there is no consensus on the age of screening, the testing method, or who – from a school nurse to a pediatrician to an eye-related provider – should administer the screening. Since delays in detection and repair of RDs are known to translate to worse functional outcomes^{4, 23}, the findings of this study may indicate that improvements could be made in universal screening using telemedicine or fundus imaging^{24, 25} and more reliable vision screening in school-age children. Innovations in childhood vision screening such as devices that prompt accurate monocular vision testing that can be administered rapidly in schools or in pediatric offices would be advantageous for early detection of visual impairment. Frequency of eye health/vision screenings in schools could be increased with such devices being more readily available.

This paper is not without limitations. First, the database used in this study does not include individuals with Medicaid insurance or who were uninsured. Lack of inclusion of these patients likely results in an underestimation of time to diagnosis and repair relative to patients who are eligible for Medicaid or are uninsured and may face greater barriers to care. This may particularly affect the pediatric population, as many children are covered by Medicaid before the age of 18. Second, the analysis of claims-based data is affected by inherent limitations such as the diagnostic accuracy and physician coding preferences. However, we aimed to minimize these inaccuracies by limiting our cohort to patients with an incident RD diagnosis, at least two confirmatory RD codes, and received treatment. Furthermore, physician reimbursement is tied to proper coding and documentation which further minimizes errors in the database. Finally, we lack clinical data such as visual acuity or clinical appearance of the detachment which limits the scope of our conclusions.

Table 2: Characteristics of patients with retinal detachment under the age of 18 by presence of previous diagnosis (specifically those increasing risk for retinal detachment).

Comparison of characteristics between patients with and without previous ocular diagnosis

		With Pre-existing Diagnosis		Without Pre-existing Diagnosis
		n/median	%/IQR	n/median	%/IQR	p-value
Number of Patients		313	37.9%	513	62.1%
Sex						0.269
	Male	202	64.5%	310	60.4%
	Female	111	35.5%	203	39.6%
Age at RD diagnosis		14	10–16	15	12–17	<0.001
Type of RD						0.295
	Rhegmatogenous	233	74.4%	403	78.6%
	Serous	18	5.8%	30	5.8%
	Tractional	62	19.8%	80	15.6%
Term status*						0.004
	Preterm	146	46.6%	186	36.3%
	Term	167	53.4%	327	63.7%
Treatment						0.295
Pneumatic retinopexy		10	3.2%	16	3.1%
Scleral buckle		90	28.8%	170	33.1%
Pars plana vitrectomy with or without scleral buckle		174	55.6%	258	50.3%
Laser barricade		35	11.2%	62	12.1%
Cryotherapy		4	1.3%	7	1.4%
Proportion with preceding eye-related visit		136	43.5%	190	37.0%	0.079
Proportion with at least two preceding eye-related visits		124	39.6%	149	29.0%	0.002
Time from last eye-related visit to RD diagnosis, days		14.5	2–140	44.5	5–214	0.011
Time from second to last eye-related visit to RD diagnosis, days		70.5	19–307	159	32–465	0.005
Time from RD diagnosis to RD repair, days		1	0–8	3	0–9	0.003

^*Preterm was defined as patients born at a gestational age of less than 37 weeks

Abbreviations:

RD

Retinal Detachment

RRD

Rhegmatogenous Retinal Detachment

VA

Visual Acuity

PVR

Proliferative Vitreoretinopathy

ICD

International Classification of Diseases

CPT

Current Procedural Terminology

IQR

Interquartile Range