Newborn Screening for Inherited Metabolic Diseases Using Tandem Mass Spectrometry in China: outcome and cost-utility analysis

Background Few studies in China have focused on economic evaluation of newborn screening(NBS) for inherited metabolic diseases (IMDs) by tandem mass spectrometry (MS/MS). This study assessed the total costs, benets, benet-cost ratio(BCR), cost-utility ratio(CUR) and incremental cost-utility ratio(ICUR) of NBS using MS/MS compared to non-screening group for the rst time. Methods This study was conducted as a retrospective piece. Newborns who underwent MS/MS screening for IMDs from 2009 to 2018 were included. All records were extracted from a screening management system in NBS Center of Zhejiang province. All costs, including indirect cost, were discounted at a rate of 5% for the whole life-time. The putative benet of clinical outcomes related to early-diagnosis through screening was assumed as improvements in quality of life and potential life expectancy in screening group. Of the 3,040,815 newborns screened, 26,297(0.86%) newborns were suspected positive after rst-round screening and called back to take another MS/MS screening. 25,670(97.62%) of them followed the latter procedures and nally 735(2.86%) cases were diagnosed through gene sequence analysis. The most frequent cause of IMD was amino acid diseases(n=276), in most cases fatty acid oxidation disorders (n=248), followed by organic acidemias (n=211). The difference of QALYs ranged from 0.78-15.4. The CUR was CNY¥86,155.80/QALY in screening group and CNY¥303,9517.32/QALY in non-screening group. The ICUR was CNY¥-795,686.47/QALY, and the BCR was 1:8.11. Neonatal intrahepatic cholestasis caused by citrin deciency; CIT-I, Citrullinemia type I; OTCD, Ornithine transcarbamylase deciency; CPS I, Carbamoyl phosphate synthetase type I; ARG, Argininemia; ASA, Arginosuccinic aciduria; HCY, Homocystinuria; MET, Hypermethioninemia; MSUD, maple syrup urine disease; TYR, Tyrosinemia; QALY, quality-adjusted life-years.


Background
Newborn screening (NBS) programs are widely recognized as secondary prevention interventions in the " eld of Public Health" around the world (1), and encompass an abundance of aspects. The program for the Early Detection of Inherited Metabolic Diseases (IMD) is part of the NBS program. IMD refers to a group of diseases with a series of clinical symptoms caused by enzyme defect, abnormal cell membrane function or receptor defect caused by gene mutation, resulting in collective biochemical metabolism disorder, and accumulation of intermediate or bypass metabolites or lack of nal metabolites (2). The screening aims to detect asymptomatic infants before onset of clinical symptoms caused by abnormal biochemical metabolism and provide proper treatment to prevent the occurrence of body dysfunction and irreversible damage, so as to reduce the morbidity, mortality and potential disabilities associated with IMD. The history of NBS began in the early 1960s, dominant with single disease screening for a long time. Methods such as gas chromatography mass spectrometry and liquid chromatography mass spectrometry also demonstrate obvious aws, mainly have long analysis cycle, low sample ux and high false-positive rate, thus cannot meet the needs of mass application for NBS. Tandem Mass Spectrometry (MS/MS) emerges as a rapid analysis technology with high sensitivity and speci city, which can test a variety of IMD including amino acids diseases, organic acidemias and fatty acid oxidation disorders through one experiment (3). The introduction of MS/MS makes possible the simultaneous measurement of several metabolites and, consequently, the detection of several diseases in one blood spot and in a unique analysis.
In China, exploratory NBS programs for IMD began with the screening of phenylketonuria (PKU) in the 1980s in Beijing and Shanghai. In 1994, the "Maternal and Infant Health Care Law of the People's Republic of China" was promulgated. For the rst time, the law proposed "gradually carry out newborn screening", which provided a fundamental legal guarantee for the work of NBS. The management measures for NBS was issued in 2009, which clearly de ned the responsibilities of health administrative departments, all levels' NBS centers and medical institutions (4). According to the data of the national However, as the expensive screening expenses and high false-positive rates, an argument has emerged that MS/MS may not be cost-effective for NBS, but should be only used as a clinical diagnostic tool for suspected positive infants. Hence, the cost-utility of using MS/MS in NBS remains to be investigated. This study rst assessed the costs, bene ts, BCR, CUR and ICUR of NBS program using MS/MS in China.

Study design
This study conducted the cost-utility analysis using a decision model. As a decision node (□), a screening or non-screening option must be selected and at a chance node (○), a variety of outcomes may occur, each one with some degree of probability. The clinical outcomes were assumed to have a difference between patients diagnosed early and late. And the costs were calculated for screening and non-screening group as the schematic diagram displayed (Fig. 1). As the falsenegative rate of MS/MS cannot be obtained in the management system, we adopted a false-negative rate of 0 in this study.
We took a societal perspective in the calculation of costs and bene ts. Incidence rate, false-positive rate and internal cost data were obtained from the NBS Management Information System of Zhejiang province, an automated system, which currently manages metabolic disorders in more than 10,000 children. Apart from that, our research team also included t pediatricians for the estimation on the frequency of follow-up and hospitalizations.
Calculations were based on average costs of medical procedures in 3,040,815 newborns who underwent MS/MS screening for IMDs in NBS center of Zhejiang province from 2009 to 2018. For the infants participated the MS/MS screening, they were asked to take the rst-round screening using dried blood spot(DBS) no sooner than age 72 hours and for premature babies, no later than age 30 days after birth, those who had positive test results were called back to take another MS/MS test. If someone had two positive test results, the con rmatory test was required to be done. Positive results from MS/MS tests would be con rmed with gene sequence analysis, genetic counselling and disease-speci ed tests for diagnosis.

IMD incidence and false-positive rates
The diseases incidence and false-positive rates were obtained from the NBS management information system of Zhejiang province, which covered the whole medical procedures of each NBS participant. For all the disorders can be detected ,28 diseases were found in the 3,040,815 newborns who underwent MS/MS screening. For costs calculation, we assumed the incidence rates and false-positive rates as the NBS management information system provided.

Estimated costs
The direct medical cost (e.g. screening tests, con rmatory tests, etc.), direct non-medical costs (e.g. special care, transportation) and indirect costs (e.g. wage loss of family member for screening) are speci ed in Table 1. On the basis of our two pediatricians' experience and Khneisser's work (5), the cost of dietary treatments and pharmaceutical treatments were not included, as they are supposed to be equally incurred once diagnosed. The cost data for screening tests, con rmatory tests, pediatrician consult, hospitalizations were extracted from the medical department.
We assumed that every newborn takes their rst-round screening in 72 hours after birth, so the rst-round MS/MS requires no extra transportation cost and wage loss. Once having a positive result, these newborns are called back to take the second-round MS/MS, which obviously need the relatives accompanied. We assumed that each newborn has two accompanying relatives and the screening make them to have a half-day away from work. Wage loss of accompanying relative was calculated by per capital income of Zhejiang province (6). In line with Zhejiang practice, the con rmatory test is required when someone gets two positive results. The normal procedure for con rmatory test necessitates a speci ed test for amino acid diseases(CNY¥60), speci ed tests for fatty acid oxidation disorders / organic acidemias (CNY¥115), a pediatrician counselling (CNY¥14), a genetic counselling(CNY¥80), and for some highly suspected cases a gene sequence test(CNY¥3,120). Most of the time, the false-positive results engender a pediatrician counselling and a speci ed test for diseases, costing as much as CNY¥74 to CNY¥129 total.
As the complications and severity are varied in different patients, some of them may have hospitalization cost. Expert opinion combined with medical records were used to estimate the hospitalization frequency of different disease. We assumed that patients with VLCAD, CPTII, OTCD, ASA, ARG, MMA, PA, IVA or Ga-may become symptomatic within the rst week of life and may require hospitalizations due to metabolic disorders in the rst 5 years. Due to less intensive hospital stays, the cost of each hospitalization for the early-diagnosed infants is CNY¥5,000. We assumed an initial hospitalization cost savings of CNY¥5,000 for early-diagnosed infants versus diagnosis made after symptoms manifested. Estimation of transportation cost during the life years of IMD patients was based on a survey conducted on 307 patients both from local and non-local regions. As the rst round MS/MS screening was tested in the local hospital, the transportation cost for this section was calculated using CNY¥18 per case. The proportion of the local and non-local con rmed cases was adopted to estimate the distributions of the newborns who were recalled. CNY¥338 per visit was used to estimate the transportation cost of non-local patients.
We use per capital basic rehabilitative services subsidy to calculate the cost of rehabilitative services (7), and the data of per capita compulsory education expenditure of disabled children to calculate the cost of special education (8).
For indirect cost, wage loss of accompanying relative /family member was calculated by per capital income of Zhejiang province. Among handicapped children, 70% can be regarded as total loss of labor force, and 30% of them can be regarded as partial loss of labor force (9). We use the gap between capital minimum income and per capital income to calculate the partial wage loss of disabled patients. For cases total loss of labor force, per capital income of Zhejiang province was used for calculation. We also assumed a 50% wage loss of one family member for each disabled patient, for in china, most of the disabled people were cared at home by the family caregivers (10,11).
As lack of discount could encourage policymakers to delay implementing health programs inde nitely (12,13). In this study, all the costs and bene ts were discounted at 5% during the life years of patients with IMDs, a commonly used gure that recommended by China Guidelines for Pharmacoeconomic Evaluations (14).

Estimated life expectancy and QALYs
Based on facts and clinical experience, the onset of complications of the primary metabolic disease in both screening patients and non-screening patients depends on the integrate operation of many factors, the putative bene t of diagnosed early through MS/MS screening was assumed as a difference in the life expectancy and QALYs after treatment. As the severity and responsiveness to treatment are varied by different disease, some even would not necessarily be detected without MS/MS screening (15), so speci ed life expectancy for each IMDs was used for calculation. We assumed a 0.5-55years difference in life expectancy between patients diagnosed late and those diagnosed early by screening. Meanwhile, the patient's burden of having an IMD had also been described by impairment in utilities (16,17). Though in some studies (18,19), adults IMD patients had similar QoL scores as normal. However, as the research design and sampling method ruled, only patients who could nish the questionnaires or reply the emails (which apparently indicates a higher cognitive level) were included leading to inevitable bias in the results. Hence, utility scores of patients with IMDs in this paper were estimated to have a 0.3-0.7 difference between early and late diagnosed patients varied by different clinical outcomes and severity of disorders. Also, a rate of 5% was adopted to calculate the total discounted QALYs as well.

Incidence of IMDs
As infants with IMDs may die without clear diagnoses, we assumed that the incidence rates in non-screening group was similar to those in screening group. A total of 3,040,815 newborns were screened between 2009 and 2018, 26,297 of them were suspected positive after rst-round screening and called back to take another MS/MS test. 25,670 of them followed the latter procedures and nally 735 cases were diagnosed through con rmatory tests. The most frequent cause of IMDs is amino acid oxidation disorders, in most cases fatty acid diseases, followed by organic acidemias. The number of diagnosed cases and incidence rates of 28 IMDs are speci ed in Table 2 (Table 4).

Utilities
In addition to, the prolongation of life years and improvement of utilities resulted in total discounted QALYs difference ranging from 0.78 to 15.40. Among them, patients with BKD and HCS seems to bene t a lot due to diagnosed early by screening. By contrast, GA-and PA could only save little QALYs by screening. The estimated life expectancy and QALYs for patients diagnosed early and diagnosed late are speci ed in Table 5 (20-.

Cost-utility analysis
Results of cost-utility analysis are presented in Table 6. The CUR was CNY¥86,155.80/QALY and CNY¥303,9517.32/QALY in screening and non-screening group. The ICUR for screening group was CNY¥-795,686.47/QALY compared to non-screening group. The costs and bene ts per patient diagnosed were CNY¥1,427,601.59 and CNY¥11,580,560.99, and the BCR was 1:8.11.

Discussion
The spectrum and incidences of IMDs differ among populations, which has been well characterized in Caucasians but much less so in Chinese (30). Several studies have been carried out in other countries (31)(32)(33)(34)(35)(36), while a growing number of studies of IMDs in Chinese population are emerging (37)(38)(39)(40)(41), mostly conducted by screening program. This study reported incidence rates of IMDs in east Chinese population using the data from NBS center of Zhejiang province during 2008Q1-2019Q4, with a total of 3 million newborns screened. The collective incidence rate of 28 IMDs included in the screening program is 1/4,137, which was lower than the reported incidence of a pilot study conducted in Mainland China (1/3795) (39) and in many other countries, e.g. 1/2500 in Canada (31), 1/2900 in Germany (32), 1/1944 in Egypt (33), 1/2916 in Malaysia (34), 1/2800 in South Korea (35), and 1/3165 in Singapore (36). Compare to other regions in China (40)(41)(42)(43), the incidence rate of IMDs in Zhejiang province also remained relatively low, excluding Taiwan (1/5882).
In this study, the most frequent cause of IMDs was amino acid oxidation disorders, in which PKU accounted for 58.33% suggesting PKU was the most common amino acid oxidation disorder and IMD in Zhejiang province. The second most common disorder was PCD, which accounted for 136 patients in this study, with its incidence rate of 1/22,359. Meanwhile, the incidence of BKD, HMG, HCS, CPS1, ASA, HCY, TYR were relatively rare, even lower than 1/1,000,000.
Early diagnosis and regular follow-up is important to improve quality of life and long-term survival of IMD, and to prevent the occurrence of body dysfunction and irreversible damage. NBS for these disorders could lighten the social burden related to incapacitating symptoms of these diseases. Though MS/MS has been conducted in China for over 10 years, study focused on the economic evaluation of this technique is still lacked, with more efforts devoted to studies on screening for single disease (e.g. CH, PKU) (44)(45).
In this study, the total cost for per early-diagnosed patient was CNY¥1,427,601.59, saving CNY¥10,152,959.41 compared to late-diagnosed patient. Savings are primarily attributed to wage loss of family members and patients (86.03%), and to avoided costs of special care (13.84%), which both bene t from the mortality and severity reduction of IMDs. The treatment cost savings only accounted for 0.08% of all the cost savings, indicating the direct medical cost would not show a substantial decrease because of MS/MS screening. Some diseases will continue to have high mortality and poor clinical outcome even detected by MS/MS before symptoms onset and well-managed since then. Table 3 presented the life-time discounted treatment cost savings of each disorder both per patient and per 100,000 newborns, the latter taking the incidence rate into consideration. It seems that MCCD, MMA, MET could each get a treatment cost saving of more than CNY¥10,000 per 100,000 newborns screened. While not considering the incidence rate, HMG,OTCD and CPT could each save more than CNY¥20,000 treatment cost per early diagnosed patient. Like some studies, we quanti ed the monetary value of expectancy life years saved and the wage loss of caregivers due to sick children as indirect cost. Failure to add indirect cost in a cost-utility analysis might make the results more likely to be not cost-effective.
The estimated cost per QALY (CUR) was CNY¥86,155.80 (for screening group) and CNY¥303,9517.32 (for non-screening group), saving CNY¥2,953,361.52 per QALY. The incremental cost per QALY (ICUR) was CNY¥-795,686.47 compared to nonscreening group, indicating NBS program could both gain QALYs and save costs at the same time.
NBS has been found to be cost-effective by other research groups in China (42)(43)(44)(45), although the estimated BCR varied widely. In their 2011 study (45), Zhang et al. reported that the estimated BCR of screening for PKU and CH in china was 1:6.9, whereas the BCRs were estimated ranging from 1:1.17 to 1:141.96 in studies carried out in each speci c disorder (43,44,46). While our study included cost-utility estimates of screening for most of the disorder categories detected by MS/MS, it did not include analyses for CH, CAH and G6PD. Inclusion of CH may have in uenced the results of the study because it is a much more prevalent condition as compared with many other disorders (47).Meanwhile, some of the bene ts of early detection of IMDs cannot be estimated through a monetary analysis, such as improved psychosocial development both in patients themselves and the caregivers. The results still showed that MS/MS screening for IMDs may be considered to be economically bene cial, which consistent with the studies carried out in other countries (5,31,(47)(48).
Despite MS/MS has been used in China for many years, its application is mainly concentrated in Shanghai, Zhejiang, Guangdong and other developed areas and high-end laboratories (49). Neither corresponding technical speci cations in China nor NBS data from less-developed areas has been reported, where might have a rate of consanguineous marriages and could lead to high incidence of IMDs. From early 2015, the NHS Newborn Blood Spot Screening Program in England started to offer screening for a total of nine disorders, six of which were IMDs. The UK screening program was regulated by the Department of Health through the National Screening Committee (UK NSC) and clear recommendations exist regarding management and follow-up of positive screen results (50). Similar technical speci cations or recommendations need to be built in China, so does the reimbursement policy related to NBS expenses nationwide.
The major limitation of this study referred to not applying a Markov model-based analysis to evaluate the costs and health bene ts for MS/MS screening in China. Moreover, this study did not consider the poor adherence of patients follow up, assuming that all patients would treat their disease for life. However, actually a substantial part of patients discontinued treatment as a result of economic problems by their parents (49).

Conclusions
In conclusion, NBS using MS/MS can be considered as cost-effective, as it provides a BCR of 1:8.11, compares favorably with some screening programs for single disease. The CUR and ICUR for screening group is CNY¥86,155.80/QALY and CNY¥-795,686. 47