Cost-Effectiveness of Treatment for Threshold Retinopathy of Prematurity

Gary C. Brown, MD, MBA*‡; Melissa M. Brown, MN, MD, MBA*§; Sanjay Sharma, MD, MSc(Epid)*࿣; William Tasman, MD‡; and Heidi C. Brown*

ABSTRACT. Objective. Retinopathy of prematurity etinopathy of prematurity (ROP) is the leading (ROP) is a leading cause of adverse visual outcomes in cause of blindness among premature infants.1 premature infants. Both laser photocoagulation and cryo- It occurs in some form in ϳ32% of infants with therapy have been demonstrated in clinical trials to be R Յ 1 birth weights 1000 g. Among those infants with efficacious in reducing the incidence of visual loss occur- birth weights Ͻ1251 g, 7% to 8% will develop thresh- ring secondary to threshold ROP. Visual data recently old ROP.2 Threshold ROP is defined as proliferative have become available concerning the long-term clinical retinal vascular disease that is located in zone 1 or 2 efficacy of both treatments, as have data concerning the in conjunction with plus disease and associated with utility value of visual states in general. Accordingly, we 5 contiguous or 8 cumulative clock hours of extrareti- undertook an analysis to ascertain the cost-effectiveness nal neovascularization.1 Of those who are treated for of laser photocoagulation and cryotherapy in the treat- threshold disease, the long-term survival is ϳ88%.3 ment of threshold ROP. The Multicenter Trial of Cryotherapy for ROP ini- Design. A computer simulation economic model is presented to evaluate the cost-effectiveness of cryother- tially demonstrated that retinal cryotherapy for threshold ROP decreased the incidence of an unfa- apy and laser photocoagulation therapy, compared with 2 the natural course of the disease, for treating premature vorable anatomic outcome from 51% to 31%. Subse- 4 infants with threshold ROP. The model applies long- quently, the Laser ROP Study Group showed with a term visual data from previous clinical trials, utility anal- meta-analysis that the risk of an unfavorable out- ysis, decision analysis, and economic principles, such as come in eyes with threshold ROP was 19% with present value analysis, to account for the time value of cryotherapy and 8% with laser therapy. money to arrive at a cost per quality-adjusted life-year Now that children who were treated for threshold (QALY) gained. ROP are maturing, data concerning the long-term Outcome Measures. Cost per QALY gained from laser visual acuity results from prospective clinical trials therapy and cryotherapy. are available.5–7 It seems that the mean visual acuity Results. Laser photocoagulation therapy for thresh- in eyes treated for threshold ROP may be better in old ROP costs $678 1998 US dollars (at a 3% discount rate those children treated with laser photocoagulation to account for the time value of money) for each QALY than in those treated with cryotherapy. Although gained from treatment. Cryotherapy for the same disease systemic complications can be associated with cryo- costs $1801 per QALY at a similar discount rate. therapy8 and laser therapy,9 the incidence of such Conclusions. From the point of view of cost-effective- complications seems to be low. ness, laser therapy seems to have an advantage over Cost-effective analyses that involve utility theory cryotherapy for the treatment of threshold ROP. are being reported with increasing frequency.10–13 Pediatrics 1999;104(4). URL: http://www.pediatrics.org/ Utility analysis incorporates patient-based prefer- cgi/content/full/104/4/e47; threshold retinopathy of pre- ences that allow the quality of life associated with a maturity, laser therapy, cryotherapy, cost-effectiveness. health (disease) state to be quantified. In view of recent data concerning utility values in patients with ABBREVIATIONS. ROP, retinopathy of prematurity; HCFA, visual loss14 and new information on visual results Health Care Financing Agency; CPT, Current Procedural Termi- after therapy for ROP,5–7 we undertook an analysis of nology; QALY, quality-adjusted life year; $/QALY, cost per qual- the cost-effectiveness of laser photocoagulation and ity-adjusted life-year. cryotherapy for the treatment of threshold ROP.

METHODS Overview From the *Center for Evidence-Based Healthcare, Flourtown, Pennsylvania; Software produced by Treeage, Inc (Williamstown, MA)15 was the ‡Retina Service and the §Cataract and Primary Eye Care Service, Wills used to perform decision analysis to simulate the clinical situation Eye Hospital, Jefferson Medical College, Philadelphia, Pennsylvania; and involving laser therapy, cryotherapy, and no therapy for patients the ࿣Epidemiology Unit, Massachusetts Eye and Ear Infirmary, Harvard with threshold ROP. This microcomputer model takes into ac- Medical School, Boston, Massachusetts. count the mortality within this group of patients, as well as the Received for publication Dec 31, 1998; accepted Apr 26, 1999. treatment results expressed in utility form. The numbers of uni- Reprint requests to (G.C.B.) Center for Evidence-Based Healthcare, Suite lateral and bilateral cases also were factored into the analysis. 210, 1107 Bethlehem Pike, Flourtown, PA 19031. E-mail: [email protected] Utility values were used at the terminal nodes at the right side of PEDIATRICS (ISSN 0031 4005). Copyright © 1999 by the American Acad- the decision tree to provide the most probable utility value out- emy of Pediatrics. comes given the different possible clinical scenarios. The utility

http://www.pediatrics.org/cgi/content/full/104/4/Downloaded from www.aappublications.org/newse47 byPEDIATRICS guest on September Vol. 27, 104 2021 No. 4 October 1999 1of6 values were based on the visual acuity in the better seeing eye old state to threshold disease at 6 months after birth, which is with lower levels of visual acuity yielding correspondingly lower when the data were reported, it was assumed that none of the utility values.14 unilateral cases would progress to bilateral threshold disease. The clinical data used for the analysis are described below. The patients with unilateral threshold ROP were presumed to Additional clinical data and assumptions used in the analysis are have a normal eye with normal vision on the unaffected side. shown in Table 1. Thus, their vision was converted to a utility value by the method shown in the next two sections. Treatment Per the criteria of the Multicenter Trial of Cryotherapy for Visual Results ROP,2 data were included in the present analysis only if the Visual results are emerging from clinical trials involving long- information involved patients with threshold ROP1 who were term follow-up of patients treated for threshold ROP.5,6 The most treated within 72 hours of the discovery of the threshold disease. specific results for treated eyes come from the data of Connolly et Cryotherapy was performed by confluently freezing all avascular al,6 who were able to quantify vision in Snellen form in threshold retina anterior to the mesenchymal ridge or extraretinal neovas- ROP eyes treated with laser therapy or cryotherapy and with a cularization in affected eyes.1,2,6 In a similar manner, only those mean follow-up of 5.8 years. The most complete visual data on the eyes receiving laser in which all avascular anterior retina was natural course of untreated threshold ROP come from the Cryo- treated with full scatter photocoagulation1,4,16 were included. therapy for ROP Cooperative Group7 that gathered data on sur- vivors from a cohort of 291 patients with threshold ROP. Mortality and Morbidity The assumption was made that once children were Ն5 years of The Cryotherapy for ROP Cooperative Study Group followed a age, the visual acuity would be stable in each eye. The incidence of cohort of 291 infants with birth weights Ͻ1251 g who entered a long-term complications associated with ROP is unknown. It was prospective randomized clinical trial and were treated in one eye assumed that the long-term complications (after 5 years of life) are with cryotherapy for threshold ROP.2,3 Data obtained on the same equal in eyes that underwent laser therapy, cryotherapy, or no treatment. cohort at 31⁄2 years after randomization revealed that there were 256 survivors among the original cohort of 291 infants.3 The av- The visual results for both treated eyes and untreated eyes were erage infant was randomized and treated at ϳ3 months of age. reported using the Snellen method, which is the most commonly used method for measuring visual acuity in clinical practice. For Thus, the mortality rate within 33⁄4 years after birth was 12% (35 the purposes of this report, the Snellen visual acuity results in the deaths of 291 infants). ϭ ϭ It was assumed that children who survived to this stage had a eyes were converted to decimal form: 20/20 1; 20/25 0.8; 20/30 ϭ 0.67; 20/40 ϭ 0.5; 20/50 ϭ 0.4; 20/60 ϭ 0.66; 20/70 ϭ normal life expectancy, although confirmatory data are not avail- ϭ ϭ ϭ ϭ able. Data on life expectancy were obtained using 1994 statistics 0.29; 20/80 0.25; 20/100 0.2; 20/200 0.1; and 20/400 0.05. obtained from the Centers for Disease Control and Prevention and For a vision of counting fingers, a value of 0.025 was assigned, and the National Center for Health Statistics.17 for hand motions, a value of 0.0125 was assigned. No light per- ception was given a value of 0.0. The incidence of postoperative retinal detachment, cataract, or 7 other ocular complications attributable to cryotherapy or laser For the natural course of the disease, those eyes classified in therapy is unknown but is likely very low. The systemic morbidity the 20/40 or better range were assigned a mean acuity of 0.67 associated with the treatments is low but present. Brown and (20/30), those in the 20/40 to 20/60 range were assigned an acuity associates8 found that cryotherapy induced cardiorespiratory ar- of 0.4 (20/50), those in the 20/60 to 20/200 range were given a rest in 1 of 80 consecutive patients, and Vander and associates9 mean acuity of 0.2 (20/100), those with 20/200 or worse were noted approximately the same results for laser therapy. assigned an acuity of 0.05 (20/400), and those who had vision that was not quantifiable were given a mean acuity of 0.0125 (hand motions). Bilaterality Versus Unilaterality of Threshold Disease For those few patients with laser or cryotherapy treatment6 The Cryotherapy for ROP Cryotherapy Group found that 82.5% who did not have specific Snellen vision measurement, a value of of the patients in the cohort of 291 infants developed bilateral 0.025 (counting fingers) was assigned to fix-and-follow or cen- threshold disease. For the purposes of our report, those patients tered, steady, and maintained vision, whereas the one eye in the with bilateral disease were assumed theoretically to have a similar series with unsteady and unmaintained vision was assigned a visual result if they were to be treated with laser therapy in both value of 0.0125 (hand motions). eyes, have cryotherapy in both eyes, or have no treatment in both eyes. Because it is rare for ROP eyes to convert from a subthresh- Conversion of Visual Acuity to Utility Values It has been noted that the mean utility values of individuals TABLE 1. Clinical Assumptions of the ROP Therapy Econo- with ocular disease diminish in proportion to the severity of visual metric Model of Cost-Effectiveness loss in the better seeing eye.14 Recent data derived via multivariate regression analysis come from a large series of patients with visual 1. Additional cryotherapy is needed in 6.4% of cases and loss from various ocular diseases.14 These data have allowed the additional laser therapy is required in the same number. conversion of Snellen visual acuity in the better eye to a mean Nevertheless, because it is performed within a 30- to 90-day utility value. The formula derived by this analysis is: postoperative period (depending upon the insurer), there is no charge by the treating physician.2 Utility ϭ 0.374x ϩ 0.514.14 2. There is no mortality from either laser therapy or cryotherapy.1,8,9 Medical Costs 3. One of 80 patients who undergoes cryotherapy or laser The costs obtained from this analysis were representative of therapy will experience a transient cardiorespiratory arrest8,9 those paid by the Health Care Financing Agency (HCFA) for that does not change the utility outcome in decision analysis; provider services classified according to Current Procedural Ter- it only impacts upon costs. minology (CPT) data.18 The data utilized were from the state of 4. Those cases with bilateral threshold retinopathy of Pennsylvania.19 The costs are summarized in Table 2. prematurity are considered to have a similar treatment result The costs themselves included the mean expense to the payer of in each eye. treatment with laser photocoagulation therapy (CPT code 67228) 5. Those with unilateral threshold ROP are presumed to have or cryotherapy (CPT code 67227) in both eyes for 82.5% of patients, normal vision (20/20) in the opposite eye. In this instance, and treatment with laser photocoagulation therapy or cryotherapy the utility is based upon the formula found for patients with in one eye in 17.5% of patients. These percentages directly reflect normal vision in one eye and visual loss to at least 20/40 in the percentages of premature infants in the Multicenter Trial of the second eye.14 Cryotherapy for ROP Study with bilateral and unilateral threshold 6. Preservation of vision is assumed to be as valuable in disease, respectively.2 The cost of an initial hospital consultation children with neurologic deficits associated with prematurity was taken from the same HCFA data using CPT code 99254. as in those with normal psychomotor and social development Because the laser therapy and cryotherapy codes most commonly parameters. have a 3-month postoperative period, no costs were attributed to

2of6 COST-EFFECTIVENESSDownloaded OF from TREATMENT www.aappublications.org/news FOR THRESHOLD by guest RETINOPATHY on September 27, 2021 OF PREMATURITY TABLE 2. Cost Data Associated with Treatment of Threshold of the safest investments (a United States government 1-year ROP treasury bill) at the time of the preparation of this article was 4.5% per year,23 and the current inflation rate in the United States was 1. The cost of laser therapy for threshold ROP is $794 per eye. It ϳ1.48% per year.24 Thus, the yearly discount rate or return on is calculated according to CPT code 67228 [destruction of investment in real dollars (calculated according to the Fisher ef- extensive or progressive retinopathy (e.g., diabetic fect) used in our analysis is 1.045/1.0148 ϭ 1.03, or 3% per year. retinopathy); one or more sessions: photocoagulation]. 2. The cost of laser therapy per patient is the weighted average RESULTS of 82.5% of patients treated in both eyes (2 ϫ $794) and 17.5% of patients treated in one eye (1 ϫ $794).2 Thus, the Utilities weighted average cost per patient is $1450.70. 3. The cost of cryotherapy for threshold ROP is $498 per eye. It The mean visual results expressed in decimal is calculated according to CPT code 67227 [destruction of equivalents using the Snellen system for acuity in extensive or progressive retinopathy (e.g., diabetic premature infants with birth weights Ͻ1251 g who retinopathy); one or more sessions: cryotherapy]. develop threshold ROP are shown in Table 3. The 4. The cost of cryotherapy per patient is the weighted average average eye with laser photocoagulation therapy has of 82.5% of patients treated in both eyes (2 ϫ $498) and 17.5% of patients treated in one eye (1 ϫ $498).2 Thus, the a mean acuity of 0.49 at 5.8 years after treatment, weighted average cost per patient is $908.85. whereas the average eye after cryotherapy has a 5. The cost of the initial examination to decide upon treatment mean acuity of 0.31 at a similar time after therapy.6 (either laser therapy or cryopexy) is CPT code 99254 (Initial Untreated eyes with threshold ROP have a mean inpatient consultation ϭ $140). 7 6. The per diem cost of an intensive care hospital bed is $1200. acuity of 0.23 at approximately the same time (5.5 7. There is no cost attributable to the follow-up examinations years after treatment). after treatment because of the 3 month postoperative period When the Snellen decimal equivalents were con- for laser therapy and cryotherapy. verted to utility values, the results were as follows: 8. The improvement in quality of life is discounted at a rate of the mean utility for patients in the laser-treated 3% over the remaining life expectancy of the long-term surviving infants who were treated ϳ3 months after birth. group was 0.70, the mean for those in the cryother- apy group was 0.63, and the mean utility for the threshold retinopathy group that received no treat- ment was 0.60. this postoperative period, despite the fact that retreatment might In the patients with threshold ROP in one eye, the have been necessary in some cases within the first few weeks of initial treatment. Additional laser photocoagulation or cryother- mean utility was 0.89. This was calculated by insert- apy for threshold ROP after this 3-month postoperative period is ing the Snellen decimal visual acuity for 20/20 or 1.0 virtually nonexistent.1 into the Utility ϭ 0.374x ϩ 0.514 formula.14 Incremental hospital costs are not associated routinely with the treatment of ROP, because the treatment now is performed rou- Decision Analysis tinely under local anesthesia,1 and the patients are already typi- cally inpatients because of the risks of marked immaturity in the Three arms were used in the decision analysis tree nonhospital setting. Nonetheless, when transient cardiorespira- for dealing with threshold ROP: 1) treatment with tory arrest occurred secondary to either cryotherapy or laser ther- photocoagulation laser therapy; 2) treatment with apy, the assumption was made that this would add 1 day to an cryotherapy; and 3) no therapy. The assumptions infant’s stay in an intensive care bed. The expense for the day was calculated from a survey of the average per diem cost for an were made that the long-term survival in each of the intensive care hospital bed in Pennsylvania. arms was 88%, and that 82.5% of cases had bilateral Anesthesia services were utilized for a percentage of patients threshold ROP. enrolled in the Multicenter Trial of Cryotherapy for ROP.2 Nev- Taking into account the above variables with the ertheless, since that time, patients are treated routinely without the aid of general anesthesia.1 For this reason, the cost of anesthe- utility values described in the previous section ap- sia services was not factored into the analysis. plied to the terminal nodes of the decision analysis tree, the mean decision analysis utility value for a Present Value Analysis patient with threshold ROP treated with laser ther- The benefits derived from laser therapy and cryotherapy for apy was 0.65. The mean utility value for a patient threshold ROP are spread out over the lifetime of the treated treated with cryotherapy was 0.59, and the mean individual, but the cost required to deliver laser therapy or cryo- utility value for patients in the group without treat- therapy is delivered during the first year of life (typically within the first 3 months of life). Therefore, some compensatory mecha- ment was 0.57. nism must be instituted to validate the economic analysis. The time value of money is not constant.20,21 A dollar spent now QALYs can be reinvested to yield more real dollars in the future. A dollar The number of QALYs is calculated by multiply- spent in the year 1999 will purchase more goods and services than a dollar spent in the year 2009. To account for the changing time ing the improvement in utility conferred by a treat- value of money, we discounted the quality-adjusted life-years ment modality by the number of expected years the (QALYs) obtained yearly from the therapy over the life expectancy modality will be in effect. For the 88% of children of the average individual to their present value. Alternatively, the who survived to the 31⁄2-year level, the mean total life future value of the dollars spent at the time of therapy also could expectancy is 76.4 years.17 Because treatment was have been calculated over the life of the patient (leaving the initial QALYs derived a constant) to arrive at the same analysis. administered at 3 months of life, and because stabi- The discount rate is variable depending on the discretion of investors or other interested stakeholders. For the purpose of this analysis, we chose to use a discount rate representative of the TABLE 3. Visual Outcome (in Snellen Decimal Equivalent) of yearly return on funds that an investor could obtain from the an Eye with Threshold ROP safest investment of capital over and above the yearly inflation Laser therapy6 0.49 Mean of 5.8 years after treatment level. In essence, this is accounting for the Fisher effect,22 which Cryotherapy6 0.31 Mean of 5.8 years after treatment adjusts the nominal return on investment for inflation, thereby No treatment7 0.23 Mean of 5.5 years after treatment yielding the real return. The approximate nominal return on one

Downloaded from www.aappublications.org/newshttp://www.pediatrics.org/cgi/content/full/104/4/ by guest on September 27, 2021 e47 3of6 TABLE 4. QALYs Gained and Cost-Effectiveness of Therapy inpatient consult fee and the cardiorespiratory arrest for Threshold ROP assessment of $15, the total cost for cryotherapy for Yearly Cryotherapy Laser Therapy the average infant was $1063. Discount Rate (%) N of QALYs $/QALY N of QALYs $/QALY Cost per QALY ($/QALY) 0 1.52 699 6.08 264 The nondiscounted US dollars expended per 3 0.59 1801 2.37 678 QALY for cryotherapy were $699 ($1063/1.52 QALYs), and for laser therapy they were $264 ($1607/6.08 QALYs). Nevertheless, these figures do lization or progression occurs within weeks at most not take into account the time value of money. Thus, after treatment, the treatment effect was assumed to a discount rate must be applied to QALYs. When the last for a total of 76 years. discount rate of 3% per year is applied to the QALYs Overall, cryotherapy conferred a ϩ0.02 utility over the 76-year period of life expectancy of the value treatment effect (0.59–0.57) per year over no average surviving infant, the $/QALY for cryother- treatment. Therefore, the number of nondiscounted, apy is $1801 and for laser therapy is $678 (Table 4). QALYs obtained by cryotherapy over no treatment Sensitivity analysis for $/QALY at a discount rate for threshold ROP was 1.52 (0.02 ϫ 76 years). of 1% per year yields a $/QALY for cryotherapy of Laser therapy conferred a ϩ0.08 utility value treat- $1003 and for laser therapy of $377. With a discount ment effect (0.65–0.57) over no treatment. Thus, the rate of 5% per year, the $/QALY for cryotherapy is corresponding gain in nondiscounted QALYs for this $2658 and for laser therapy is $1017. With a large therapy over no treatment was 6.08 (0.07 ϫ 76 years). discount rate of 10% per year, the corresponding As mentioned in a previous section, the number of figures are $5315 for cryotherapy and $2030 for laser QALYs obtained by a treatment regimen need to be therapy. discounted to provide a valid economic analysis. Employing a 3% discount rate over 76 years to the DISCUSSION QALYs obtained with cryotherapy, the number of The results of this study suggest that both cryo- discounted QALYs obtained from cryotherapy over therapy and laser therapy are exceedingly cost-effec- the natural course of the disease was 0.59 (Table 4). tive treatments for threshold ROP. Both treatments The number of QALYs obtained from laser therapy result in a substantial improvement in mean quality over the natural course of the disease was 2.37. of life that lasts over the lifetime of the average Sensitivity analysis of the values using varying patient with threshold ROP. Although this study discount rates revealed that with a 1% discount rate noted that the actual patient treatment costs associ- over the 76-year benefit period, cryotherapy deliv- ated with cryotherapy ($1063) are less than those ered 1.06 QALYs and laser therapy delivered 4.26 associated with laser therapy ($1607), evidence from QALYs. When the discount rate was increased to 5% a randomized clinical trial6 has shown that eyes for the 76-year period of benefit, the numbers of treated with laser therapy have better long-term QALYs gained by cryotherapy and laser therapy visual results than those treated with cryotherapy. were 0.40 and 1.58, respectively. With a substantially These recent results were not completely unexpected higher discount rate of 10% over the 76 years, cryo- by those well versed in treating threshold ROP. Re- therapy showed a gain of 0.20 QALY, and laser ther- sults from the Laser ROP Study Group4 previously apy showed a gain of 0.79 QALY. suggested the increased clinical efficacy of laser ther- apy over cryotherapy from the anatomic point of Medical Costs view. The cost of an initial inpatient consult to examine Overall, using a 3% discount rate over the 76-year a threshold child was $140 for both cryotherapy- and period of life expectancy of a patient with threshold laser-treated patients. Assuming that 1 of 80 cryo- ROP, the cost for laser therapy is $678/QALY and therapy- or laser therapy-treated patients would re- the cost for cryotherapy is $1801/QALY. The greater quire an extra day in intensive care attributable to the long-term clinical efficacy of laser therapy and the sequelae of transient cardiorespiratory arrest, the av- subsequent increased improvement of quality of life erage cost per patient at a per diem hospital day conferred by the better visual results suggest that it ($1200 per day) was $15. should be considered to be the preferred treatment if The cost of laser therapy for one threshold eye was a clinician must choose between laser therapy and $794, and the cost for two eyes was $1592. Thus, the cryotherapy. Nevertheless, in instances with cloudy weighted average cost of laser therapy per patient media (such as from vitreous hemorrhage or a (assuming bilateral threshold disease in 82.5% of marked tunica vasculosa lentis), laser therapy may threshold cases) was $1452. With the additional $140 not be possible, and cryotherapy may be the only for consultation and $15 for the risk of cardiorespi- option. ratory arrest, the total expense related to the treat- When comparing the cost-effectiveness of one ment was $1607 for the average infant. medical interventional treatment with another, care The cost of cryotherapy for one eye was $498, and must be taken to minimize the number of confound- for two eyes was $996. Using the same weighted ing factors.25 If possible, utility valuation should be average that takes into account the 82.5% of bilateral undertaken with the same investigational methodol- cases, the mean cost for cryotherapy of a threshold ogy (using the time trade-off method, the standard infant was $908. Therefore, with the additional $140 gamble method, or another method),26,27 because the

4of6 COST-EFFECTIVENESSDownloaded OF from TREATMENT www.aappublications.org/news FOR THRESHOLD by guest RETINOPATHY on September 27, 2021 OF PREMATURITY results of these different methods can differ substan- therapy, whereas the present study concentrated ba- tially.28,29 Utility values obtained from different pop- sically on interventional treatment for threshold ulations (eg, patients, physicians, administrators, disease. Neither the Javitt paper nor this paper took and the general public) also can differ substantially into account the long-term follow-up costs associated for the same disease state.27,30 We believe, as do oth- with ROP over a lifetime. More importantly, how- ers,31–33 that, when possible, preferences obtained ever, was the fact that long-term visual acuity data from patients should be used, because those who after cryotherapy and laser treatment for threshold have experienced diseases firsthand are best able to ROP were not available for Javitt and colleagues.37 assess the degree to which the disease impairs qual- They assumed that good and bad anatomic fundus ity of life. The utility values used in the present study results correlated with either excellent vision or were derived solely from patient preferences.14 Sim- blindness, respectively, which is not the case.7 Thus, ilar discounting methods and rates also should be the utility assumptions were substantially different used to account for differences in the time value of in that paper, compared with the present paper. De- money caused by inflation, opportunity costs, or cur- spite the differences, both the Javitt paper and the rency exchange differences. The sensitivity analysis present study demonstrate the substantial cost-effec- in the present study demonstrates the vast differ- tiveness of treatment for ROP. ences in results that can occur from altering the The present study has theoretical drawbacks that discount rate alone, particularly when the treatment should be noted. The study by Connolly et al,6 on benefit effect occurs over many years and/or in the which the follow-up visual acuities were based, had distant future. These substantial differences reflect a relatively small cohort of patients, and the fol- the effect of compounding that occurs when calcu- low-up period was only 5.8 years; the visual sequelae lating discount rates. Finally, comparing medical of laser therapy and/or cryotherapy after 5.8 years costs should reflect a uniform scale, such as the scale are not known. Additionally, the utility values that used by the HCFA.19 Comparing cost-efficiency from we used were calculated from a formula derived by one country to another is particularly difficult be- multivariate analysis from a large sample of adults cause of the lack of standard costs and the effect of with visual loss.14 It can be argued that, because currency exchange fluctuations.34 visual loss from ROP typically occurs earlier in life, Keeping the above inexactitudes in mind, for com- these patients might adapt better and have a higher parison we have calculated that a single vessel cor- utility score associated with visual loss than do onary artery bypass procedure for left main coronary adults who lose vision later in life. Nonetheless, it artery disease is associated with an incremental, 1998 has been demonstrated in a large study that there is US dollar cost-effectiveness of $6880/QALY30 and not a correlation between the length of time of visual that a liver transplant is associated with an incremen- loss and utility values.14 tal cost-effectiveness of $327 500/QALY.36 To our Although an overestimation utility impairment in knowledge, data concerning the incremental cost- ROP patients may bias the results toward decreased effectiveness of most interventional procedures asso- cost-effectiveness, the visual data used in our study ciated with the sequelae of prematurity are not pres- may actually underestimate the cost-effectiveness of ently available. Although there are no absolute treatment for threshold disease. The mean age of standards for cost-effectiveness, it has been sug- patients from who the visual data were obtained was gested arbitrarily that interventional therapies cost- Ͻ6 years. The difficulty in coaxing children of that ing Ͼ$100 000/QALY are not particularly cost-effec- age with perfect eyes to read beyond the 20/30 or tive, whereas those costing Ͻ$20 000/QALY are 20/25 line is well appreciated by ophthalmologists. cost-effective.35,36 Therefore, it is likely that if the visual results were Thus, both laser therapy and cryotherapy for ROP measured 5 years later, the mean vision would be seem to be excellent values. The permanent improve- improved at the upper end of the scale. ment in vision that translates to improved quality of It should be noted that premature children who life from these treatments, which are of moderate develop neurologic impairment were included in the cost, explains why they are especially cost-effective. economic analysis in the present study. It has been In essence, laser therapy and cryotherapy are clini- estimated that ϳ22% of infants with a birth weight cally efficacious treatments with long-standing ben- Ͻ1250 g have a major handicap with neurologic eficial effects that are not prohibitively expensive by involvement.38 Other authors37 have excluded this modern health care standards. group in cost-effectiveness analysis for ROP, but we Previous authors have studied the cost-effective- believe that vision is just as valuable to those with ness of screening and cryotherapy for threshold neurologic impairment, and perhaps even more so, ROP. Javitt and associates37 wrote an excellent trea- as to those without neurologic impairment. tise on the subject in 1993 and noted that appropriate Also of note is the fact that our analysis did not ophthalmic screening of premature infants and sub- include Markov modeling in the decision analysis sequent cryotherapy treatment for threshold disease aspect. Markov modeling is a useful tool when a produce a substantial overall saving for the United decision analysis involves risk that is continuous States at a cost of $2488 to $6045/QALY, depending over time.39 Although the software program that we on the screening strategy. There are several impor- used has substantial Markov process capability,15 the tant differences between that paper and the present incidence of long-term complications associated with study. Javitt and associates37 took into account gen- ROP are unknown. Because assumptions are often eral ROP-screening costs in addition to those of cryo- the most important part of economic modeling, we

Downloaded from www.aappublications.org/newshttp://www.pediatrics.org/cgi/content/full/104/4/ by guest on September 27, 2021 e47 5of6 chose to ignore this issue rather than use subjective ulation for stage 3ϩ retinopathy of prematurity. Ophthalmology. 1991; and potentially misleading information. Nonethe- 98:576–580 17. US Department of Health and Human Services. Vital Statistics of United less, we believe that the incidence of serious ocular States, 1994. Preprint of Volume II, Mortality, Part A, Section 6. Hyatts- complications (eg, retinal detachment, cataract, and ville, MD: US Department of Health and Human Services; 1998. DHHS 1 glaucoma) after 53⁄4 years of age is relatively low and Publication No PHS 98-1104 that the difference in the late ocular complication 18. Kirschner CG, Davis SJ, Duffy T, et al. Physicians’ Current Procedural rates between those who are treated for ROP and Terminology 98. Chicago, IL: American Medical Association; 1998:20–24, those who are not is likely to be minimal. 255–268 19. Relative Value Studies, Inc. The Complete RBRVS. Reston, VA: St An- thony Publishers; 1998:23–443 ACKNOWLEDGMENTS 20. Weinstein MC, Stasson WB. Foundations of cost-effectiveness analysis This work was supported in part by the Retina Research and for health and medical practices. N Engl J Med. 1977;296:716–721 Development Foundation, Philadelphia, PA, and by the Princi- 21. Damodaran A. Corporate Finance: Theory and Practice. New York, NY: pal’s Initiative and Research Fund, Kingston, Ontario, Canada. John Wiley & Sons, Inc; 1997:37–65 22. Ross SA, Westerfield RW, Jordan BD. Fundamentals of Corporate Finance. REFERENCES 3rd ed. Chicago, IL: Irwin; 1995:112–165 1. McNamara JA, Connolly BP. Retinopathy of prematurity. In: Regillo 23. Bonds. In: Barron’s: The Dow Jones Business and Financial Weekly. Novem- CD, Brown GC, Flynn HW Jr. Vitreoretinal Disease: The Essentials. New ber 30, 1998:MW 60 York, NY: Thieme; 1999:177–192 24. US Department of Labor Bureau of Labor Statistics. Consumer Price 2. Cryotherapy for Retinopathy of Prematurity Cooperative Study Group. Index. Washington, DC: US Department of Labor Bureau of Labor Multicenter trial of cryotherapy for retinopathy of prematurity: three Statistics. Available at: ftp://ftp.bls.gov/pub/special.requests/cpi/ month outcome. Arch Ophthalmol. 1990;108:195–204 cpiai.txt. Accessed October 18, 1998 3. Cryotherapy for Retinopathy of Prematurity Cooperative Group. Mul- 25. Siegel JE, Weinstein MC, Russell LB, Gold MR. Panel on cost- ticenter trial of cryotherapy for retinopathy of prematurity: 31⁄2-year effectiveness in health and medicine. JAMA. 1996;276:1339–1341 outcome—structure and function. Arch Ophthalmol. 1993;111:339–344 26. Redelmeier DA, Detsky AS. A clinician’s guide to utility measurement. 4. The Laser ROP Study Group. Laser therapy for retinopathy of prema- Med Decis Making. 1995;22:271–280 turity. Arch Ophthalmol. 1994;112:154–156 27. Brown MM, Brown GC, Sharma S, Garrett S. Evidence-based medicine, 5. White JE, Repka MX. Randomized comparison of diode laser photoco- utilities and quality of life. Curr Opin Ophthalmol. 1999;10:221–226 agulation versus cryotherapy for threshold retinopathy of prematurity: 28. Wakker P, Stiggelbout A. Explaining distortions in utility elicitation 3-year outcome. J Pediatr Ophthalmol Strabismus. 1997;34:83–87 through the rank-dependent model for risky choices. Med Decis Making. 6. Connolly BP, McNamara JA, Sharma S, Regillo CD, Tasman W. A 1995;15:180–186 comparison of laser photocoagulation with trans-scleral cryotherapy in 29. Richardson J. Cost utility analysis: what should be measured? Soc Sci the treatment of threshold retinopathy of prematurity. Ophthalmology. Med. 1994;39:7–21 1998;105:1628–1631 30. Torrance GW, Feeny D. Utilities and quality-adjusted life years. Int J 7. Cryotherapy for Retinopathy of Prematurity Cooperative Group. Mul- Technol Assess Health Care. 1989;2:559–575 ticenter trial for retinopathy of prematurity: Snellen visual acuity and 31. Kassirer JP. Adding insult to injury: usurping patients’ prerogatives. 1 structural outcome at 5 ⁄2 years after randomization. Arch Ophthalmol. N Engl J Med. 1983;308:898–901 1996;114:417–424 32. Angell M. Patients’ preferences in randomized clinical trials. N Engl 8. Brown GC, Tasman WS, Naidoff M, Schaffer DB, Quinn G, Bhutani VK. J Med. 1984;310:1385–1387 Systemic complications associated with retinal cryoablation for retinop- 33. Kassirer JP. Incorporating patients’ preferences into medical decisions. athy of prematurity. Ophthalmology. 1990;97:855–858 N Engl J Med. 1994;330:1895–1986 9. Vander JF, Handa J, McNamara JA, et al. Early treatment of posterior 34. Drummond MF. Economic aspects of cataract. Ophthalmology. 1988;95: retinopathy of prematurity: a controlled clinical trial. Ophthalmology. 1147–1153 1997;104:1731–1736 35. Detsky AS. Are clinical trials a cost-effective investment? JAMA. 1989; 10. Torrance GW. Measurement of health state utilities for economic ap- 262:1795–1800 praisal. A review. J Health Econ. 1986;5:1–30 11. Torrance GW. Utility approach to measuring health-related quality of 36. Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a life. J Chron Dis. 1987;40:593–600 new technology have to be to warrant adoption and utilization: tenta- 12. Lee TT, Solomon NA, Heidenreich PA, Oehlert J, Garber AM. Cost- tive guidelines for using clinical and economic evaluation. Can Med effectiveness of screening for carotid stenosis in asymptomatic patients. Assoc J. 1992;146:473–481 Ann Intern Med. 1997;126:337–346 37. Javitt J, Dei Cas R, Chiang YP. Cost-effectiveness of screening and 13. Brown GC, Sharma S, Brown MM, Garrett S, and the Quality in Med- cryotherapy for threshold retinopathy of prematurity. Pediatrics. 1993; icine Study Research Group. Evidenced-based medicine and cost- 91:859–866 effectiveness. J Health Care Finn. In press 38. Thompson CM, Buccimazza SS, Webster J, Malan AF, Molteno CD. 14. Sharma S, Brown GC, Brown MM, Snow K, Sharma SM, Brown H. Infants less than 1250 grams birth weight at Groote Schuur Hospital: Converting visual acuity to utilities. Can J Ophthalmol. In press outcome at 1 and 2 years of age. Pediatrics. 1993;91:961–968 15. Data 3.0 User’s Manual. Williamstown, MA: Treeage Software, Inc; 1997 39. Sonnenberg FA, Beck JR. Markov models in medical decision making: a 16. McNamara JA, Tasman W, Brown GC, Federamn JL. Laser photocoag- practical guide. Med Decis Making. 1993;13:322–338

6of6 COST-EFFECTIVENESSDownloaded OF from TREATMENT www.aappublications.org/news FOR THRESHOLD by guest RETINOPATHY on September 27, 2021 OF PREMATURITY Cost-Effectiveness of Treatment for Threshold Retinopathy of Prematurity Gary C. Brown, Melissa M. Brown, Sanjay Sharma, William Tasman and Heidi C. Brown Pediatrics 1999;104;e47 DOI: 10.1542/peds.104.4.e47

Updated Information & including high resolution figures, can be found at: Services http://pediatrics.aappublications.org/content/104/4/e47 References This article cites 28 articles, 3 of which you can access for free at: http://pediatrics.aappublications.org/content/104/4/e47#BIBL Subspecialty Collections This article, along with others on similar topics, appears in the following collection(s): Neonatology http://www.aappublications.org/cgi/collection/neonatology_sub Ophthalmology http://www.aappublications.org/cgi/collection/ophthalmology_sub Health Alerts http://www.aappublications.org/cgi/collection/health_alerts Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.aappublications.org/site/misc/Permissions.xhtml Reprints Information about ordering reprints can be found online: http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on September 27, 2021 Cost-Effectiveness of Treatment for Threshold Retinopathy of Prematurity Gary C. Brown, Melissa M. Brown, Sanjay Sharma, William Tasman and Heidi C. Brown Pediatrics 1999;104;e47 DOI: 10.1542/peds.104.4.e47

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://pediatrics.aappublications.org/content/104/4/e47

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since 1948. Pediatrics is owned, published, and trademarked by the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 1999 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news by guest on September 27, 2021