To determine the cost-effectiveness of laser treatment for retinopathy of prematurity (ROP) in Lima, Peru.
A cost-of-illness study (in US dollars) to determine the direct cost of treatment, the indirect lifetime cost of blindness, and the quality-adjusted life years.
The direct cost of ROP-related treatment was determined by reviewing data retrospectively from a social security sector hospital. The indirect cost was determined using national economic data of Peru published by the Central Information Agency (CIA), including the per capita gross domestic product, the sex-adjusted income distribution, and years spent in the work force. Indirect costs per child that were avoided by treatment were calculated using the known natural history of ROP vs evidence-based treatment.
For ROP-related neonatal blindness in Peru, we estimate the total indirect cost saving at $197 753 per child and the direct cost of laser treatment at $2496 per child. The societal lifetime cost saving per child is estimated at $195 257. The mean annual income per educated adult in Peru is $8000 and treating 1 child is equivalent to employing 24 educated Peruvians per year. The generational cost savings for society is approximately $516 million, or the equivalent of 64 500 educated Peruvian work years.
The societal burden of blindness far exceeds the costs of treatment per child. Proper screening and treatment of ROP prevents blindness and leads to substantial cost savings for society. Public health policy in Peru and other middle-income countries should consider financial impact when allocating healthcare resources.
Retinopathy of prematurity (ROP) is rapidly becoming a common and alarming cause of neonatal blindness in the developing world. Significant improvements in neonatal care are leading to an increase in survival rates of premature and extremely premature neonates with ROP. The World Health Organization (WHO) defines blindness as a best-corrected visual acuity in the better eye of less than 20/400 and low vision as best-corrected visual acuity in the better eye of worse than 20/60 but equal to or better than 20/400. Using this definition, the prevalence of blindness in children (0–15 years) worldwide is estimated to be 1.4 million. Data suggest that in highly industrialized countries (ie, ranked by the United Nations Development Programme on the basis of their Human Development Index), the infant mortality rate is very low at <9/1000 live births, with the rate of ROP blindness in this population at 10%. In countries with high infant mortality rates (above 60/1000), the rate of ROP-related blindness is extremely low, likely because of the lack of neonatal facilities and a higher mortality rate. On the other hand, countries with infant mortality rates between 9/1000 and 60/1000 have a very high and concerning rate of ROP-related blindness. This demographic of increasingly at-risk neonates is being referred to as the “third epidemic” of ROP. The majority of countries classified as middle-income are areas in Latin America, Eastern Europe, Southeast Asia, and the Caribbean. In Latin America, approximately 24% of childhood blindness occurs secondary to ROP, with an estimated 24 000 affected children. In 2002, a multicenter study was published using data from 11 neonatal intensive care units in Latin America. The authors identified a 42% rate of ROP in very low-birthweight infants, with only 68% of the babies receiving adequate screening.
Peru, a middle-income country, has a population of 29 million and a childhood blindness prevalence of 0.062%, with approximately 17 000 blind persons <15 years old. A survey in 2003 conducted at 4 schools for the blind in Lima determined that the frequency of ROP-related blindness was 16%. The same survey performed 2 years later in 2005 reported an increased frequency of ROP-related blindness, to 25% (L Gordillo, personal communication, Lima, Peru, September 19, 2010).
This increasing frequency of ROP deserves further investigation, with consideration of improved public policy support for clinical activities, because few ophthalmologists are available to screen and treat ROP. In 2002 there were only 800 ophthalmologists in Peru, but the number was expected to grow by 50 per year. At present, there are few primary ophthalmologists who conduct ROP screening examinations; therefore, imaging and fundus photographic screening technologies are currently being implemented as these forms of screening improve and begin to gain acceptance.
Based on ROP studies conducted in the United States, significant reductions in the rate of blindness from ROP have been demonstrated, initially with the use of cryotherapy and later with the indirect diode laser–delivered panretinal photocoagulation (current standard treatment). The long-term outcome data from the Cryotherapy for Retinopathy of Prematurity study demonstrated that 15 years after treatment, the number of unfavorable structural outcomes was 52% in the control group in comparison to 30% in the treatment group.
In Peru, current treatment uses the diode laser. The solid-state diode lasers are easy to transport and allow access for treatment into more rural areas. In a series of consecutive eyes with threshold ROP treated using diode laser photocoagulation, the reported rate of regression among 8 studies ranged from 71% to 100%. In addition, the Early Treatment of Retinopathy of Prematurity study reported that laser treatment for high-risk prethreshold ROP reduced the rate of unfavorable structural outcomes from 15.6% to 9.0%. Effective evidence-based reduction in the incidence of blindness from ROP emphasizes the importance of early detection, screening, and laser treatment of prethreshold ROP.
The medical infrastructure and neonatal care continues to improve in Peru and other middle-income nations. Along with such improvements, a greater burden of ROP is expected. The societal impact of ROP and subsequent childhood blindness must be considered in order to appropriately allocate scarce healthcare resources in the most cost-effective manner. Traditionally, a cost-of-illness study is used to address such conditions. These studies take into account both the direct and indirect costs, including productivity loss secondary to morbidity or premature mortality. Direct costs include the cost of screening, medical staff, treatment, and follow-up examinations. Indirect cost includes the individual’s lost productivity plus that of the caretaker. Caretakers of blind adults are estimated to lose at least 10% of their productivity, clearly an additional burden on society. To the best of our knowledge, there are no current estimates for the lost productivity for caretakers of blind children. Overestimates of cost may include future tax revenue, a cost that is not included in this study and may falsely overestimate the impact. Instead, we propose that a more fact-based analysis should be used.
The purpose of this work is to demonstrate that evidence-based ROP care is cost-effective in Peru, reduces the societal burden of blindness, and substantially improves the quality of life for those directly and indirectly affected. Furthermore, this analysis will also help the political entities in such countries to best determine appropriate and cost-effective allocation of scarce healthcare resources. The financial and societal analysis taken from this study may also be applicable to other developing countries. This dataset will demonstrate and highlight the magnitude and cost-effectiveness of aggressive screening and prevention strategies for management of treatable neonatal blindness from ROP. Such a model analysis could be applied to many developing nations.
In 2002, Resnikoff and associates presented data on the prevalence of blindness in Peru. By combining their data with published data from 4 blind schools, we estimated the frequency of blindness from ROP in 2005.
With the assistance of our coauthor (L.G.), a de-identified database was reported from previously collected data at the Hospital Nacional Edgardo Rebagliati Martins EsSalud neonatal intensive care unit (NICU) over the first 6 months of 2010 in order to determine the number of neonates who were screened for ROP. Stages of ROP (1–5) either with or without plus disease were determined at each examination, as well as the number of infants that were treated. There are 47 NICUs in Peru, with 24 located in Lima. An additional de-identified database from 2009 for 11 NICUs in Lima was used to determine the most current estimates of babies treated per year. Local ethical standards for data collection were followed and conducted upon consultation with our collaborator (L.G.).
The average direct medical cost to treat ROP was calculated from data acquired from the Hospital Nacional Edgardo Rebagliati Martins EsSalud NICU. These costs include 3 major components: labor (physician, nursing, and medical staff), equipment and supplies (indirect ophthalmoscope, examination lens, sterile lid speculum, and dilating drops), and facility (estimated additional time and space allocation for ROP-related care). All were determined using available data from the local hospital NICU during a site visit in 2010.
The direct treatment costs per child were determined based on amortized equipment costs over 8 years and included available equipment-related service contracts. Nine neonates were treated during a 6-month period of record review (9 × 2 × 8 = 144). As the incidence of ROP is increasing, we estimate that approximately 160 treatments will be required in this NICU over the next 8 years. We estimate an 8-year life span of the laser, with limitations to further use attributable to transportation-related wear, plus possible future evolution of the technology. The cost to treat a generation, 20 years, was estimated taking into consideration the lifetime of the equipment and estimating that 400 children would be treated over this time frame.
The indirect cost of blindness to a given society is calculated as the sum of total years of lost productivity and opportunity cost of informal care given by the family member who cares for the blind individual. The gross domestic product (GDP) per capita in international units (provided by the Central Intelligence Agency World Factbook) was used to determine the societal costs related to lost productivity of a blind individual. Lost productivity per capita was used to determine lifetime productivity loss with a sex-adjusted estimate from the Socio-Economic Database for Latin America and the Caribbean (SEDLAC). This productivity shift assumed that every reasonably sighted individual over age 18 (legal working age in Peru) would contribute to the GDP. We used the average, sex-adjusted retirement age and calculated the loss in productivity of a blind individual as defined by the WHO. We adjusted our estimates for ROP-related blindness on estimates of current treatment based on the Early Treatment of Retinopathy of Prematurity outcome data and compared these to the natural history group in the Cryotherapy for Retinopathy of Prematurity study, with the 15-year outcome data serving as the best evidence-based data set for natural history outcomes of untreated threshold ROP in comparison to the outcomes influenced by current treatment standards. Therefore, we estimate the lifetime GDP lost per child treated in 2010 at the Hospital Nacional Edgardo Rebagliati Martins EsSalud NICU, based on 20 treated neonates (10 male, 10 female).
Next, we estimated the indirect costs associated with lost productivity of the caregiver. An average number of hours spent caring for a blind child was determined from a survey of 35 parents (mostly female) who were raising blind children in Lima. We also assumed that the amount of care per child would decrease to 3 hours per day from age 18 until death, based on reports from parents and educators at the schools for the blind. We adjusted the data to account for the mortality of blind individuals in a developing country, nearly half the mortality of sighted individuals. We calculated the total indirect cost by assessment of the lifetime sex- and age-related adjustment loss in GDP per capita plus the cost of informal care. Informal care included helping the child with daily activities such as bathing, dressing, and feeding as well as providing assistance with transportation to school and activities (specifically, additional care for these activities that was considered beyond that required for a sighted child). Finally, the total lifetime direct and indirect cost per blind child was calculated and multiplied by the frequency of ROP-related blindness statistics in order to determine the total cost to society.
Data collected in 2009 from 11 NICUs in Lima showed that 1160 babies were screened, with 71 babies (6.1%) requiring laser treatment. At the Hospital Nacional Edgardo Rebagliati Martins EsSalud NICU during the 6 month period of March – October 2010, 11% (9 of 79) of the babies that were screened needed treatment for ROP, with 1 child requiring 2 separate treatments.
For the first year (20 neonates), the monthly cost of medical staff was $894 × 12 and the monthly nursing staff was $1252 × 12, for a total annualized labor cost of $25 752. The cost of screening equipment included a 30-diopter lens ($286) and an indirect ophthalmoscope ($1073) (total 1-time cost of $1359). The 1-time cost of laser equipment was $46 512 plus $2147 annual maintenance fee (total of $48 659). Annual treatment cost for the facility was $748 per treatment, for a total of $14 960 for 20 babies. Follow-up care included an average of 3 visits per child ($54), for a total of $1080 for 20 children. The total annualized cost for this cohort was $91 810. For the subsequent years, the initial cost of the equipment would not be incurred but the cost of yearly maintenance would remain. During subsequent years, to treat 20 babies the cost of labor would be $25 752, plus $2157 for laser maintenance, $14 960 for professional fees, and $1080 for follow-up care, for a sum total of $43 949. Taking into account the depreciation of the equipment, the cost to treat a generation (20 years) would be $998 390. An estimate of 400 babies would be treated over 20 years, leading to the estimated cost of $2496 per child ( Table 1 ).
|Nursing staff||Per month||894|
|Medical staff||Per month||1252|
|30-D lens||One-time cost||286|
|Indirect light||One-time cost||1073|
|Equipment||One-time cost||46 512|
|# of follow-up visits||Average||3|
|Total||For the first year||91 810|
|For the following years (2nd–7th)||43 949|
|Total cost to treat a generation||20 years||998 390|