Journal of Pediatric Ophthalmology and Strabismus

Original Article 

Toward Achieving 100% Adherence for Retinopathy of Prematurity Screening Guidelines

Robert W. Arnold, MD; Jack Jacob, MD; Zinnia Matrix

Abstract

Purpose:

To report on the use of a cloud-based electronic medical record (ROP Check; Glacier Medical Software, Anchorage, AK) designed to provide American Academy of Pediatrics (AAP) guideline-adherent retinopathy of prematurity (ROP) care through the scheduling and documenting of ROP examinations.

Methods:

Data analysis on 3,155 patients from a de-identified dataset from 13 neonatal intensive care units.

Results:

All newborns with a gestational age of 22 to 30 weeks (N = 2,278) were entered with a documented ROP examination. Of those, 98% and 97.4% completed their initial and follow-up examinations, respectively, within AAP guidelines. All but 1 of 145 initial treatments were completed within AAP guidelines after a decision for treatment was made. Of 369 newborns older than 30 weeks' gestational age and with a birth weight of less than 1,500 grams, none progressed to treatment; four patients had stage 2 or 3 ROP. Of 508 newborns with a gestational age of 31 to 32 weeks and a birth weight of more than 1,500 grams who were entered to identify unstable newborns, 34% did not need examinations; of those who were examined, one progressed to treatment. Fourteen percent of patients were observed as outpatients before retinal maturity and 12% missed some or all outpatient appointments. A decision was made to treat at the first examination for 2 (10%) newborns with a gestational age of 22 weeks and 2 (2%) newborns with a gestational age of 23 weeks. Each patient was within the AAP guidelines for initial treatment.

Conclusions:

A computerized system specifically designed with process improvement and error-free delivery of ROP care as a focus can improve adherence to AAP guidelines and achieve superior results. Such a system creates a standardized and measurable safety net for pediatric ophthalmologists, neonatologists, hospitals, and follow-up pediatricians.

[J Pediatr Ophthalmol Strabismus. 2017;54(6):356–362.]

Abstract

Purpose:

To report on the use of a cloud-based electronic medical record (ROP Check; Glacier Medical Software, Anchorage, AK) designed to provide American Academy of Pediatrics (AAP) guideline-adherent retinopathy of prematurity (ROP) care through the scheduling and documenting of ROP examinations.

Methods:

Data analysis on 3,155 patients from a de-identified dataset from 13 neonatal intensive care units.

Results:

All newborns with a gestational age of 22 to 30 weeks (N = 2,278) were entered with a documented ROP examination. Of those, 98% and 97.4% completed their initial and follow-up examinations, respectively, within AAP guidelines. All but 1 of 145 initial treatments were completed within AAP guidelines after a decision for treatment was made. Of 369 newborns older than 30 weeks' gestational age and with a birth weight of less than 1,500 grams, none progressed to treatment; four patients had stage 2 or 3 ROP. Of 508 newborns with a gestational age of 31 to 32 weeks and a birth weight of more than 1,500 grams who were entered to identify unstable newborns, 34% did not need examinations; of those who were examined, one progressed to treatment. Fourteen percent of patients were observed as outpatients before retinal maturity and 12% missed some or all outpatient appointments. A decision was made to treat at the first examination for 2 (10%) newborns with a gestational age of 22 weeks and 2 (2%) newborns with a gestational age of 23 weeks. Each patient was within the AAP guidelines for initial treatment.

Conclusions:

A computerized system specifically designed with process improvement and error-free delivery of ROP care as a focus can improve adherence to AAP guidelines and achieve superior results. Such a system creates a standardized and measurable safety net for pediatric ophthalmologists, neonatologists, hospitals, and follow-up pediatricians.

[J Pediatr Ophthalmol Strabismus. 2017;54(6):356–362.]

Introduction

The publication of the policy statement “Screening of Premature Infants for Retinopathy of Prematurity” by the American Academy of Pediatrics (AAP), American Academy of Ophthalmology, American Association for Pediatric Ophthalmology and Strabismus, and American Association of Ophthalmologists in 2013 set the standard for the screening and treatment of retinopathy of prematurity (ROP) in the United States.1

Implementing these guidelines can be a challenge for many reasons, partially due to the complexity of the neonatal intensive care unit (NICU) setting, where pediatric ophthalmologists provide episodic ROP care. It is not uncommon for newborns in such a setting to be transferred to another NICU or to be sent home by an attending neonatologist before retinal maturity is reached and without the approval of the ophthalmologist who had been providing ROP care. ROP care is unique because the risk of missed or delayed examinations is potentially catastrophic and can lead to blindness. This makes it especially important that examinations be executed with a precision that is rarely mandatory in other health care environments.

These issues have been observed in litigation related to ROP care,2 which has become the second most common area of litigation in the NICU setting (Pediatrix Medical Group, personal communication, April 2014). ROP care litigation is generally related to missed or late examinations, poor communication between the NICU and ophthalmologist, and the transition in care between inpatient and outpatient settings.2 Therefore, it is not surprising that many pediatric ophthalmologists have opted out of providing ROP care, in which they are burdened with the risk of providing ROP care but have little input into the decisions that the NICU staff make on newborns at risk for ROP.

Most NICUs use a paper-based or spreadsheet-based system for scheduling and following up with newborns at risk for ROP. However, results have been suboptimal. An analysis in 2012 from the California Perinatal Quality Care Collaborative showed that only 77% to 83% of newborns meeting AAP screening guidelines were screened.3 This was also consistent with the Vermont Oxford Network data, which showed that 66% to 73% of newborns with a gestational age of 29 weeks or younger or a birth weight of less than 1,500 grams were screened.4 Both studies also reported a high degree of variation in screening rates between hospitals. Although these numbers were surprisingly low, they are the only data available that are based on large clinical datasets rather than a controlled setting. The reason for these low numbers is unknown and may be due to clinical practice or inaccuracies in data entry.

Neither study addressed other aspects of ROP care, such as the timeliness of screening examinations, inadvertently missed examinations, or patients lost to follow-up after discharge or transfer from a hospital.

Organizations such as the National Quality Forum and California Perinatal Quality Care Collaborative have identified adherence to the AAP guidelines as an important quality measure in NICU care.5

Clearly, there is a need to effectively solve this issue that allows for certainty of decision making in scheduling initial and follow-up ROP examinations. Traditional electronic medical records used in hospitals have not solved this issue, partially because of their general medical applications that override their focus on ROP care.

Patients and Methods

We have developed a cloud-based electronic medical record (ROP Check; Glacier Medical Software, Anchorage, AK) for scheduling, tracking, and documenting ROP examinations.6 The program has built-in decision support based on AAP guidelines that allows for the tracking of guideline adherence. Recent modifications to the program in 2015 allow for cases of infants at risk for ROP to be followed not just in their NICU of origin, but also after discharge from the NICU or transfer to another NICU.

We report on 3,155 surviving patients in 13 NICUs in the United States who have been using the program variably from 2011 through December 2015. There are 10 Level III and 3 Level IV NICUs. In this study, we report on the performance of the program as used by pediatric ophthalmologists in a clinical, day-to-day, non-research setting with an emphasis on adherence to AAP guidelines.

The program functions in conjunction with an ROP coordinator who enters all newborns with a gestational age of 32 weeks or younger and a birth weight of less than 1,500 grams into the program. Newborns with a gestational age of 30 weeks or younger are automatically scheduled for examination at 31 weeks' gestational age or 4 weeks' chronologic age, whichever occurs later, per AAP guidelines. Because the AAP has no definite recommendations for the timing of examinations for newborns older than 30 weeks' gestational age and with a birth weight of less than 1,500 grams, such newborns are scheduled at the discretion of the individual institutions and scheduling generally occurs at 2 to 4 weeks' chronological age to capture this risk group before discharge from the NICU. Similarly, newborns with a gestational age of 31 to 32 weeks who have a birth weight of more than 1,500 grams and have an unstable clinical course are also scheduled per neonatology and ophthalmology recommendations at a few weeks' chronological age based on the discretion of the individual institutions.

Generally, examinations are scheduled for a certain day of the week based on the ophthalmologist's preference. The AAP recommends that examinations be performed at 31 weeks' postmenstrual age for a newborn with a gestational age of 27 weeks or less. Adherence to AAP guidelines with a 3-day grace period would extend the acceptable examination to 32 weeks' and 2 days' postmenstrual age. For the purposes of this study, this definition is used when referring to examinations performed within a grace period.

Patients who were referred to Level IV hospitals or outpatient care for eye examinations or treatments weeks after initial ROP assessments at another institution were excluded from the analysis (217 infants).

This study is a descriptive summary of ROP care performance and practice across diverse clinical practice settings. Institutional review board approval was obtained for the study.

Results

The number of patients and examinations conducted at each gestational age is shown in Figure 1. A total of 2,278 newborns with a gestational age of 22 to 30 weeks were entered and 9,045 examinations were conducted.

Number of patients and examinations at each gestational age.

Figure 1.

Number of patients and examinations at each gestational age.

A total of 2,278 initial examinations were performed on newborns with a gestational age of 30 weeks or younger. Of those entered into ROP Check, 100% had an eye examination performed unless the patient died or was transferred to a hospital not using ROP Check. Within one of the hospitals where the authors provide ROP care and the actual number of newborns with a gestational age of 30 weeks or younger is known, 100% of such newborns had at least one examination performed unless he or she was transferred to another hospital or died (this was cross-checked against hospital electronic medical records and clinical entry logs). Ninety-eight percent of newborns with a gestational age of 30 weeks or younger had their initial eye examination performed within AAP guidelines with a grace period of 3 days. The AAP has guidelines on follow-up examinations based on examination findings. In our cohort, 97.4% of patients had follow-up examinations performed within AAP guidelines with a grace period of 1 day. No patient had missed examinations.

A total of 145 (6%) initial treatments were completed. One hundred percent of patients meeting AAP threshold criteria for treatment received treatment. All but one patient received treatment within the AAP guideline of 3 days after treatment decision; this patient received treatment 4 days after the decision for treatment was made.

We identified 5 newborns who had a treatment decision made at the time of the initial examination. Of those, 2 were 22 weeks', 2 were 23 weeks', and 1 was 26 weeks' gestational age.

All 5 newborns identified with threshold ROP were treated at 31 to 32 weeks' gestational age. None of these patients developed stage 4 or more ROP and all were treated successfully without progression.

There were 369 newborns older than 30 weeks' gestational age and with a birth weight of less than 1,500 grams admitted into ROP Check. They represented 12% of the newborns in ROP Check and 3.8% of the ROP examination burden. The characteristics of the newborns in this category are shown in Table 1. Of these newborns, 4% did not have any eye examinations and 95% did not have ROP. Two newborns had stage 2 ROP and 2 newborns had stage 3 ROP. One of the newborns reaching stage 3 died. A substantial number of patients in this category (45%) needed more than one examination. No patient needed treatment.

Characteristics of Newborns With a Gestational Age of 31 Weeks or Older and a Birth Weight of < 1,500 Grams (N = 369)

Table 1:

Characteristics of Newborns With a Gestational Age of 31 Weeks or Older and a Birth Weight of < 1,500 Grams (N = 369)

AAP guidelines state that newborns older than 30 weeks' gestational age who have an unstable clinical course may need ROP screening. In our cohort, 508 newborns entered into ROP Check were 31 to 32 weeks' gestational age and had a birth weight of 1,500 grams or more. They represented 16% of the newborns entered into ROP Check and 3.4% of the examination burden. The characteristics of the newborns in this category are shown in Table 2. Of these newborns, 34% did not need or get eye examinations and 97% did not have ROP. However, one newborn developed type 1 ROP and needed treatment.

Characteristics of Newborns With a Gestational Age of 31 to 32 Weeks and a Birth Weight of ≥ 1,500 Grams (N = 508)

Table 2:

Characteristics of Newborns With a Gestational Age of 31 to 32 Weeks and a Birth Weight of ≥ 1,500 Grams (N = 508)

Experience suggests that an increasing number of ROP examinations on immature retina are performed after discharge from the NICU. There have been 756 patients admitted into ROP Check since the addition of the outpatient capability in 2015 to seamlessly transfer infants to an outpatient ophthalmology setting and document examinations in ROP Check. Three of the 13 hospitals saw patients as outpatients and 102 (14%) patients had outpatient appointments. Twelve (12%) of these patients failed to keep their outpatient appointments, of which 3 patients had documentation that the provider had changed as the reason for a failed appointment. Seven (7%) of the patients seen as outpatients missed one or more appointments. Fifty-one percent of patients seen as outpatients had a gestational age of 22 to 26 weeks, 38% had a gestational age of 27 to 30 weeks, and 11% had a gestational age of older than 30 weeks.

Of the outpatients, 19% and 9% had stage 2 and 3 ROP, respectively. Three patients needed treatment, two as inpatients and one as an outpatient after discharge from the NICU. Thirty-nine percent of patients were younger than 40 weeks' postmenstrual age when they were first seen as outpatients.

Discussion

Many pediatric ophthalmologists do not use traditional electronic medical records because the non-specific software fails to meet the specialized needs of ROP care. Many NICUs use either a paper-based or spreadsheet-based system for screening and tracking newborns.

Ophthalmic Mutual Insurance Company, the insurer for many ophthalmologists in the United States, has developed a paper-based “safety net” that outlines best practices for ophthalmologists caring for ROP.7 Such systems, although important, are prone to human error, take considerable time to implement, and are frequently dependent on a single individual in the NICU or ophthalmologist's office performing the task. Furthermore, the results produced by such systems are unknown. A retrospective review of large clinical databases, such as the Vermont Oxford Network and California Perinatal Quality Care Collaborative, and our past deliberate review of our paper-based scheduling reveal that not all newborns with a gestational age of 30 weeks or younger are examined despite concerted efforts by ophthalmologists to ensure that every at-risk patient gets examined in a timely manner. It is unknown whether there are untoward consequences of these missed examinations. California Perinatal Quality Care Collaborative data suggest that many of the newborns not observed are of more advanced gestational ages.3 In our series, 3.5% of newborns with a gestational age of 28 to 30 weeks developed stage 3 ROP and 2.8% needed treatment. This suggests that excluding low-risk newborns poses a small but significant risk for an adverse outcome.

Through ROP litigation cases, we also know that litigation is frequently related to process issues such as missed or delayed examinations and missteps in the transition of care either between NICUs or between the NICU and home.2 Traditional electronic medical records have not addressed this issue and pediatric ophthalmologists have frequently opted out of direct interaction with such systems. In addition, because some ophthalmologists have opted out of ROP care, others have assumed responsibility for providing care at multiple hospitals, making the tracking of patients in different settings increasingly cumbersome and prone to error. ROP Check was developed as a solution that addresses these issues in ROP care.6

We report on the performance of a system designed to optimize the process of providing ROP care in both simple and complex clinical ROP practices. This is the first study to document routine ROP clinical practices across a diverse set of hospitals in the United States with an emphasis on adherence to AAP guidelines. The only information available on clinical ROP practice through the Vermont Oxford Network and California Perinatal Quality Care Collaborative datasets indicates that a small but significant percent of newborns meeting screening criteria are not screened.3,4 Our study shows that a computerized system specifically designed with process improvement and error-free delivery of care as a focus appears to improve adherence to AAP guidelines. Such a system, especially when used in conjunction with a ROP coordinator, has the ability to schedule 100% of newborns needing examinations in a timely manner.

This is the first study to track adherence to all AAP guidelines in a clinical setting. We were able to screen 100% of newborns with a gestational age of 30 weeks or younger in a diverse clinical setting. This assumes that all newborns with a gestational age of 30 weeks or younger were entered into ROP Check. The ability of ROP Check to capture 100% of newborns meeting criteria for ROP examinations depends on having an ROP coordinator enter all eligible newborns reliably. However, the ability exists for this process to become more reliable and automated through the pulling of defined demographic data from an electronic medical record into ROP Check. Unfortunately, our experience so far is that hospitals are reluctant to have their electronic medical records interact with other software.

The timing of initial ROP examinations occurred 98% of the time within AAP guidelines given a 3-day grace period. Subsequent examinations occurred on time, without any missed examinations 97.4% of the time. We documented almost perfect adherence to treatment guidelines for both criteria for treatment and the timing of treatment once the decision to treat was made. This creates a measurable safety net for pediatric ophthalmologists, neonatologists, hospitals, and follow-up pediatricians alike.

There are two additional categories of newborns for whom there is no information on clinical practice. First, newborns with a gestational age of 30 weeks or older who have a birth weight of less than 1,500 grams, although not needing treatment for ROP in our dataset, did have a 1% incidence of stage 2 to 3 ROP. With a larger population in a more diverse group of practices, it is possible that a few such newborns might reach threshold for treatment. This confirms the need to include such newborns in screening as suggested by the AAP. Second, newborns with a gestational age of 31 to 32 weeks and a birth weight of more than 1,500 grams who have “an unstable clinical course” are also captured by the program. The risk for ROP in such newborns is unknown. Our data indicate that such newborns do have risk, albeit low, for significant ROP, which confirms the benefit of including such newborns in screening.

Additional clinical information would need to be investigated to clarify whether a subgroup of newborns with a gestational age of 31 weeks or older carry a risk of ROP to decrease the examination burden on ophthalmologists without compromising care.

Because of a shortage of pediatric ophthalmologists and the proliferation on NICUs in the United States, attention has been given to decreasing the examination burden for ophthalmologists. Screening guidelines should capture all newborns who may need treatment. This is especially true for newborns at risk for ROP because the risks of missing or delaying a treatment are devastating. A similar situation exists with other newborn screening protocols, such as genetic newborn metabolic screens that are designed to capture all newborns affected by certain rare conditions. We are not suggesting that all units use the criteria of a gestational age of 31 to 32 weeks and a birth weight of more than 1,500 grams as screening criteria. We are simply reporting on this feature of ROP Check that is meant to capture more mature and larger preterm newborns who are at risk for ROP because of an unstable clinical course. The ROP care of such newborns has been unknown until now because there are no published studies on this category of newborns. ROP Check is the only program meant to identify such newborns so that their clinical course can be reviewed and a decision can be made regarding whether screening examinations are warranted.

AAP guidelines have a caveat stating that, in newborns with a gestational age of 25 weeks or younger, consideration should be given to start initial ROP screening examinations before 31 weeks' postmenstrual age. Our data imply that this is a reasonable recommendation; the decision was made to treat 10% and 2% of newborns with a gestational age of 22 and 23 weeks, respectively, at their first examination. However, it is reassuring that, even in this high-risk group of premature newborns, the AAP guidelines functioned as intended because each newborn with early-onset severe ROP was identified in a timely manner for treatment. ROP Check provides an earlier scheduling option for ophthalmologists and NICUs desiring earlier screening of prethreshold ROP.

Our study also identifies the issue of ROP examinations being performed after discharge from the NICU in patients who have not achieved retinal maturity. Our data indicate that this practice varies between institutions and practices. A small but substantial number of patients (19%) missed one, more than one, or all appointments. This suggests that there needs to be more focus on this practice, including the education and communication with neonatologists on the substantial risk in discharging a patient from the NICU who has not achieved retinal maturity. Coordination with ophthalmologists in decision making is critical in preventing adverse outcomes.

Our study has several limitations. First, the study may not be representative of all practices across the United States. Institutions and pediatric ophthalmologists electing to use ROP Check represent a select group that may have a different focus on ROP care. Second, detailed clinical information on patients is not available. Therefore, our ability to predict who within a category of gestational age or birth weight is at little to no risk for significant ROP is unknown. Thus, we are unable to address the issue of minimizing the workload of ophthalmologists involved in ROP care, which is an issue that has received attention in the ophthalmology literature.8 However, combining detailed clinical information within the structure of ROP Check has the ability to illuminate this important subject.

References

  1. Fierson WMAmerican Academy of Pediatrics Section on OphthalmologyAmerican Academy of OphthalmologyAmerican Association for Pediatric Ophthalmology and StrabismusAmerican Association of Certified Orthoptists. Screening examination of premature infants for retinopathy of prematurity. Pediatrics. 2013;131:189–195. doi:10.1542/peds.2012-2996 [CrossRef]
  2. Reynolds JD. Malpractice and the quality of care in retinopathy of prematurity (an American Ophthalmological Society thesis). Trans Am Ophthalmol Soc. 2007;105:461–480.
  3. Bain LC, Dudley RA, Gould JB, Lee HC. Factors associated with failure to screen newborns for retinopathy of prematurity. J Pediatr. 2012;161:819–823. doi:10.1016/j.jpeds.2012.04.020 [CrossRef]
  4. Nightingale Internet Reporting System. Vermont Oxford Network Web site. https://nightingale.vtoxford.org. Published: 2012.
  5. Perinatal and Reproductive Health 2015–2016. National Quality Forum Web site. https://www.qualityforum.org/Publications/2016/12/Perinatal_and_Reproductive_Health_2015-2016_Final_Report.aspx. Published: December, 2016.
  6. Arnold RW, Jacob J, Matrix Z. A cloud-based electronic medical record for scheduling, tracking, and documenting examinations and treatment of retinopathy of prematurity. J Pediatr Ophthalmol Strabismus. 2012;49:342–346. doi:10.3928/01913913-20120710-01 [CrossRef]
  7. OMIC ROP Safety Net. Ophthalmic Mutual Insurance Company Web site. https://www.omic.com/rop-safety-net/. Updated: April 4, 2016
  8. Hutchinson AK, Melia M, Yang MB, VanderVeen DK, Wilson LB, Lambert SR. Clinical models and algorithms for the prediction of retinopathy of prematurity: a report by the American Academy of Ophthalmology. Ophthalmology. 2016;123:804–816. doi:10.1016/j.ophtha.2015.11.003 [CrossRef]

Characteristics of Newborns With a Gestational Age of 31 Weeks or Older and a Birth Weight of < 1,500 Grams (N = 369)

Variable%
Gestational age at birth (wk)
  31 to 3276%
  33 to 3523%
  ≥ 362%
Birth weight (g)
  < 1,0006%
  1,000 to 1,24928%
  1,250 to 1,49966%
No. of examinations per newborn
  155%
  228%
  310%
> 3 (range: 4 to 10)7%

Characteristics of Newborns With a Gestational Age of 31 to 32 Weeks and a Birth Weight of ≥ 1,500 Grams (N = 508)

VariableNewborns Not Needing An ExaminationNewborns Needing an Examination
Gestational age at birth (wk)
  3167%69%
  3233%31%
Birth weight (g)
  1,500 to 1,99978%87%
  2,000 to 2,49922%11%
  ≥ 2,5000%2%
No. of examinations per newborn
  173%
  221%
  34%
  > 32%
Authors

From Alaska Children's Eye and Strabismus, Anchorage, Alaska (RWA); Alaska Neonatology Associates, Mednax Medical Group, Anchorage, Alaska (JJ); and Glacier Medical Software, Anchorage, Alaska (RWA, JJ, ZM).

Drs. Arnold and Jacob are co-owners and Ms. Matrix is a board member of Glacier Medical Software.

Correspondence: Robert W. Arnold, MD, Alaska Children's Eye and Strabismus, 3500 Latouch #280, Anchorage, AK 99508. E-mail: eyedoc@alaska.net

Received: September 15, 2016
Accepted: February 23, 2017
Posted Online: August 29, 2017

10.3928/01913913-20170329-04

Sign up to receive

Journal E-contents