The direct-medical costs associated with interferon-based treatment for Hepatitis C in Vietnam

Abstract

Background: Injectable interferon-based therapies have been used to treat hepatitis C virus (HCV) infection since 1991. International guidelines have now moved away from interferon-based therapy towards direct-acting antiviral (DAA) tablet regimens, because of their superior efficacy, excellent side-effect profiles, and ease of administration. Initially DAA drugs were prohibitively expensive for most healthcare systems. Access is now improving through the procurement of low-cost, generic DAAs acquired through voluntary licenses. However, HCV treatment costs vary widely, and many countries are struggling with DAA treatment scale-up. This is not helped by the limited cost data and economic evaluations from low- and middle-income countries to support HCV policy decisions. We conducted a detailed analysis of the costs of treating chronic HCV infection with interferon-based therapy in Vietnam. Understanding these costs is important for performing necessary economic evaluations of novel treatment strategies.

Methods: We conducted an analysis of the direct medical costs of treating HCV infection with interferon alpha (IFN) and pegylated-interferon alpha (Peg-IFN), in combination with ribavirin, from the health sector perspective at the Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam, in 2017.

Results: The total cost of the IFN treatment regimen was estimated to range between US$1,120 and US$1,962. The total cost of the Peg-IFN treatment regimen was between US$2,156 and US$5,887. Drug expenses were the biggest contributor to the total treatment cost (54-89%) and were much higher for the Peg-IFN regimen.

Conclusions: We found that treating HCV with IFN or Peg-IFN resulted in significant direct medical costs. Of concern, we found that all patients incurred substantial out-of-pocket costs, including those receiving the maximum level of support from the national health insurance programme. This cost data highlights the potential savings and importance of increased access to generic DAAs in low- and middle-income countries and will be useful within future economic evaluations.

Introduction

The World Health Organization (WHO) estimates that there are 71 million people living with chronic hepatitis C infection globally ¹. Hepatitis C virus (HCV) is typically transmitted through intravenous drug use, unsafe injection practice and the transfusion of unscreened blood/blood products ¹. There are six major HCV strains, genotypes 1–6, and the prevalence of each genotype varies significantly between regions ¹). Currently, most data relate to the treatment of genotypes 1–4; few data exist regarding the treatment outcomes and costs of genotype 6 infection, which accounts for over 50% of HCV infections in Vietnam ².

Interferon-based therapies have been used to treat HCV since 1991 ³. The original interferon (IFN) intramuscular injections had to be administered daily and were associated with poor cure rates and unpleasant side effects ⁴. Pegylated IFN (Peg-IFN) was first licensed in 2001 ⁵ and has improved pharmacokinetics, requiring only weekly injections. Additionally, Peg-IFN is more effective than IFN and is associated with fewer adverse effects ⁶. When used for 24–48 weeks with the anti-viral tablet ribavirin, Peg-IFN is associated with cure rates of 54–63% (depending on the infecting viral genotype) ⁷. In recent years, new oral direct-acting antivirals (DAAs) have been developed. DAA therapy requires a shorter duration of treatment (typically 12 weeks) and has superior cure rates to interferon-based therapy (>95%) ⁸. The tablets have almost no side effects and forgo the need for weekly injections ⁹. When they first emerged, DAA drugs were prohibitively expensive for many healthcare systems, making them unavailable in most low- and lower-middle-income countries. Whilst prices still restrict access in many settings, the situation is improving, with steep price reductions for DAAs, driven largely by increased competition from generic manufacturers and the issuing of voluntary licenses ¹⁰. In 2016, 86% of people starting HCV therapy worldwide received DAA drugs rather than interferon-based therapy ¹¹. Recently, the WHO recommended that interferon-based therapy should no longer be used where DAA drugs are available ⁸.

In 2016, the WHO released the first Global Health Sector Strategy on viral hepatitis with a goal of eliminating viral hepatitis as a public health threat by 2030 ¹². The specific goals set for hepatitis C were that 80% of patients are treated, along with a 90% reduction in the incidence and a 65% reduction in HCV related mortality ¹². In 2015, only 7% of the 71 million people living with chronic HCV infection were treated ¹². Therefore, access to treatment needs to expand if the elimination goals are to be achieved.

Although improving, the global scale-up of DAA treatment has been markedly uneven, with a handful of countries (e.g. Egypt, China) accounting for the majority of the increase in uptake ¹¹. A WHO analysis of country experiences of DAA scale-up shows that, while access to affordable treatment is important, countries also need a strong government response, including national plans for preventing, diagnosing and treating HCV, and adequate financing to roll out and sustain HCV services ^{11, 13}. For this to occur, it is vital to have a detailed understanding of the cost and cost-effectiveness of the different treatment options available in low- and middle-income countries ^{14, 15}.

We conducted a detailed analysis of the costs of treating chronic HCV with the pre-existing standard of care in Vietnam, IFN and Peg-IFN therapy. Since 2016, Vietnam has started to move away from interferon-based therapy towards DAA treatment regimens (in keeping with WHO guidelines). However, until very recently (June 2019), DAAs were not covered by national health insurance. In addition, data on the costs of interferon-based therapy are still essential for conducting accurate economic evaluations of DAA treatment, as interferon-based treatment will likely be the comparator ( Box 1) within the analysis.

Box 1. Glossary.

Catastrophic health expenditure	When the medical expenditure of a household exceeds a certain level of capacity such that the household has to cut down on necessities (such as food, clothing, and their children's education).
Comparator	Within an economic evaluation, the new intervention being investigated is compared to a comparator. The comparator generally reflects the current clinical practice.
Co-payment	The proportion of the total billed healthcare costs that insured patients pay.
Direct costs	The costs related to the goods, services and resources consumed to implement and access healthcare.
Direct medical costs	The costs directly related to the use of medical services/resources (such as physician services, diagnostic tests and drugs).
Direct non-medical costs	The costs related to the consumption of non-medical resources (such as transportation to the health facility, food expenses and accommodation).
Health sector perspective	A perspective that only includes the costs associated with the health sector, such as the costs covered by the national health insurance programme and the patient’s copayment for the medical services.
National health insurance programme	An insurance programme managed by the government that helps patients pay for medical services.
Out-of-pocket payment	The medical expenses incurred by patients that do not get reimbursed by insurance programmes.
Perspective	The viewpoint adopted for deciding which types of costs and health benefits are to be included within an economic evaluation.
Productivity costs (indirect costs)	Represent the value of the productivity losses that result from illness, treatment, or premature death.
Societal perspective	A perspective that includes all the costs associated with an intervention/healthcare, regardless of whom they are incurred by, i.e. this includes the health systems and patient’s direct medical costs, direct non-medical costs and indirect costs.
Universal health coverage	All individuals have access to good quality healthcare services without facing financial hardship.

Methods

Study location

Vietnam is a lower-middle-income country in Southeast Asia with a population of over 95 million people ¹⁶, and a 2017 gross per capita income of US$2160 ¹⁷. The seroprevalence of HCV in the general population has been estimated to be between 1 and 4.7% ^{18, 19}, which is high relative to other countries in the region. In Vietnam, genotypes 1 and 6 predominate ^{18, 20}. These genotypes are considered hardest to treat with interferon-based therapies and both genotypes 1 and 6 require prolonged treatment courses (48 weeks as opposed to 24 weeks) ²¹.

The Hospital for Tropical Diseases (HTD) in Ho Chi Minh City is the major referral hospital for infectious diseases in the south of Vietnam. Our cost estimation was performed in the context of the HTD in 2017.

The resources and services required for HCV treatment

The Ministry of Health (MoH) approved four interferon-based treatments within their first HCV treatment guidelines in 2013: IFN α-2a, IFN α-2b, Peg-IFN α-2a and Peg-IFN α-2b ²². To enhance treatment efficacy, each of these injection-based treatments is combined with the antiviral tablet ribavirin ( Table 1). In late 2016, the MoH published an updated treatment guideline for Hepatitis C, in which the recommended treatments were Peg-IFN and DAAs. Although standard IFN was no longer included as a recommended treatment, it remained on the list of medicines covered by the national health insurance programme ( Box 1). From the end of 2016, IFN was no longer used at HTD. However, as it was still used in other hospitals in Vietnam, we have included cost analysis related to IFN treatment within this paper.

Table 1. The Vietnam MoH treatment guidelines for HCV drugs.

Name of drugs	Dose
IFN α-2a	3 million IU three times per week
IFN α-2b	3 million IU three times per week
Peg-IFN α-2a	180 μg once per week
Peg-IFN α-2b	1.5 μg/kg once per week
Ribavirin	Genotype 1/4/6: 1000mg per day
	Genotype 2/3: 800mg per day

Based on the HCV guidelines from the MoH in 2013 ²². IU, international unit.

We estimated the quantity of the drugs required for an average treatment based on the recommended dosages within the 2013 HCV treatment guidelines from the MoH ( Table 1) ²². The dosage of Peg-IFN α-2b was calculated assuming an average body weight of 58 kilograms for men and 50 kilograms for women ²³.

The utilisation of the other resources and services (such as the medical tests) required to provide HCV treatment were based on the MoH 2013 treatment guidelines ²². These recommend that a patient should visit the outpatient clinic once prior to treatment, every four weeks during treatment and once after ending treatment. A summary of the required medical tests at these different stages of treatment is shown in Table 2. Following the hospital’s classification, the tests were grouped into seven different classes. The duration of treatment depends on the genotype of HCV: 24 weeks for genotypes 2/3 and 48 weeks for genotypes 1/4/6 ²². Both of these regimens were considered in our analysis.

Table 2. A summary of the recommended medical tests within the Vietnam Ministry of Health (MoH) hepatitis C virus (HCV) treatment guidelines.

Name of required tests		Before treatment	During treatment	After treatment
Group 1: Electrocardiogram	Electrocardiogram	Yes	No	No
Group 2: Ultrasound	Abdominal ultrasound		Every 12 weeks
	Fibro-scan		No
Group 3: X-ray	Chest X-ray		No
Group 4: Blood tests	Full blood count		Every 4 weeks
	The international normalized ratio		No
	Prothrombin		Every 12 weeks
Group 5:Immunoassay	Alpha-fetoprotein
	Free thyroxine
	Thyroid - Stimulating Hormone
	Hepatitis B surface antigen		No
	Human immunodeficiency virus
Group 6: Biochemical tests	Electrolytes
	Albumin
	Bilirubin
	Creatinine (urine and blood)		Every 4 weeks
	Urea		No
	Alanine transaminase		Every 4 weeks
	Aspartate aminotransferase		No
	Gamma-glutamyl transferase		No
Group 7: Molecular biology tests	HCV-RNA viral load test		Every 8 weeks	Yes
	HCV genotype real-time PCR		No	No

Before treatment, certain tests are required to assess disease severity and to ensure that treatment can be safely tolerated. During treatment, monitoring tests are required every 4, 8 or 12 weeks to assess treatment response and drug side effects. After treatment, the HCV-RNA viral load test is repeated to assess treatment response. This is based on the 2013 HCV treatment guidelines from the MoH ²².

Cost estimation and outputs

Our cost analysis estimated the direct medical cost of HCV treatment from the health sector perspective ( Box 1). The direct non-medical costs and indirect costs ( Box 1) were not quantified.

The identified resources and services ( Table 1 and Table 2) were costed based on the services and drug unit price list of HTD in 2017 relating to those covered by the national health insurance programme ²⁴. One exception to this was the costs relating to the IFN drugs which were obtained from a 2017 report from the Drug Administration of Vietnam ²⁵.

The main output was the total cost of the different treatments, stratified by the three main cost components: the cost of the drugs, the cost of the medical tests and the costs related to the clinical consultation fees.

All costs were converted to US dollars (US$) following the average 2017 exchange rate where 22,370 Vietnamese dong (VND) equal 1 US$ ²⁶.

Results

The total cost of the IFN treatment regimen was estimated to range between US$1,120 and US$1,962 and the total cost of the Peg-IFN treatment regimen between US$2,156 and US$5,887 ( Table 3). The cost of treating genotypes 1/4/6 (which require a 48-week treatment regimen) was substantially higher than the cost of treating genotypes 2/3 (which require a 24-week regimen). The cost was not exactly double, due to the different dosages of ribavirin used for the different genotypes ( Table 1 and Table 3) and the fact that the pre-and post-treatment tests are identical.

Table 3. The estimated cost of the different regimens.

Treatment	Drugs (US$)	Tests (US$)	Consultation fees (US$)	Total (US$)
Genotype 2/3: 24-week treatment regimen
IFN α-2a + ribavirin	619.17 – 655.17	498.18	13.95	1,155.29 – 1,167.31
IFN α-2b + ribavirin	607.75 – 619.77	498.18	13.95	1,119.88 – 1,131.90
Peg-IFN α-2a + ribavirin	1,644.14 – 2,617.89	498.18	13.95	2,156.27 – 3,130.02
Peg-IFN α-2b + ribavirin	1,680.07 – 1,967.78	498.18	13.95	2,192.20 – 2,470.91
Genotype 1/4/6: 48-week treatment regimen
IFN α-2a + ribavirin	1,325.52 – 1,333.78	613.24	24.41	1,953.84 – 1,961.90
IFN α-2b + ribavirin	1,254.71 – 1,262.97	613.24	24.41	1,882.83 – 1,892.09
Peg-IFN α-2a + ribavirin	3,327.48 – 5,259.20	613.24	24.41	3,955.60 – 5,887.32
Peg-IFN α-2b + ribavirin	3,375.89 – 3,950.73	613.24	24.41	4,003.43 – 4,578.85

The range in the costs for a given regimen is due to the variation in the costs of the different brands of the drugs and the different dosages (minimum and maximum values are shown in Table 3). Costs are in 2017 prices.

We estimated the costs of the three main components of HCV treatment: the drugs, medical tests and clinical consultation fees. The costs of the drugs contributed between 54–89% to the total treatment cost and were much higher for the Peg-IFN regimen. These were shown as a range because the exact price varies depending on which brand is used ( Table 4). This variation was most significant for Peg-IFN α-2a. The cost of the ribavirin only represented 2–8% of the costs relating to the drugs.

Table 4. The assumed input and unit cost for the direct medical cost based on interferon.

Item	Total Quantity		Unit cost (VND)	Unit cost (US$)
	Genotype 2/3	Genotype 1/4/6
Drug
Feronsure (3×10 6 IU)	216×10 6	432×10 6	189000	8.32
Superferon (3×10 6 IU)	216×10 6	432×10 6	178000	7.84
Pegasys (135 μg)	4320	8640	1797313	79.14
Pegasys (135 μg) new version	4320	8640	2327195	102.48
Pegasys (180 μg) old version	4320	8640	1400000	61.65
Pegasys (180 μg) new version	4320	8640	1950000	85.87
Pegnano (180 μg) old version	4320	8640	1750000	77.06
Pegnano (180 μg) new version	4320	8640	1500000	66.05
Peg-intron (50 μg) a	Man: 2088 Woman:1800	Man: 4176 Woman: 3600	1014860	44.69
Peg-intron (80 μg) a	Man: 2088 Woman:1800	Man: 4176 Woman: 3600	1639400	72.19
Peg-intron redipen (100 μg) a	Man: 2088 Woman:1800	Man: 4176 Woman: 3600	2058000	90.62
Barivir (400 mg)	403200	336000	2900	0.13
Barivir (500 mg)	403200	336000	3900	0.17
Medical tests
Group 1: Electrocardiogram
Electrocardiogram	1	1	45900	2.02
Group 2: Ultrasound
Abdominal ultrasound	3	5	49000	2.16
Fibro-scan	1	1	79500	3.50
Group 3: X-ray
Chest X-ray	1	1	69000	3.04
Group 4: Blood tests
Full blood count	1	1	44800	1.97
The international normalized ratio	1	1	12300	0.54
Prothrombin	3	5	61600	2.71
Group 5: Immunoassay
Alpha – fetoprotein	3	5	90100	3.97
Free thyroxine	3	5	63600	2.80
Thyroid -stimulating Hormone	3	5	58300	2.57
Hepatitis B surface antigen	1	1	712000	31.35
Human immunodeficiency virus	1	1	319700	14.08
Group 6: Biochemical tests
Electrolytes	1	1	28600	1.26
Albumin	1	1	21200	0.93
Bilirubin	1	1	21200	0.93
Creatinine (urine)	6	12	15900
Creatinine (blood)	7	13	21200	0.93
Urea	1	1	21200	0.93
Alanine transaminase	1	1	21200	0.93
Aspartate aminotransferase	7	13	21200	0.93
Gamma – glutamyl transferase	1	1	19000	0.84
Group 7: Molecular biology tests
HCV-RNA viral load test	5	8	1310000	57.69
HCV genotype: Real time PCR	1	1	1550000	68.25
Consulation
Clinical consultation fees	8	14	39000	1.72

^aThe dosage is prescribed following average body weight for men is 58 kg and for women is 50 kg ²⁷. Costs are in 2017 prices. IU, international unit

The costs of the medical tests were also notable (US$498 for treating genotypes 2/3 and US$613 for treating genotypes 1/4/6) ( Figure 1). The HCV-RNA viral load tests accounted for the majority of this (approximately 70%). The costs relating to the tests used pre- and post-treatment were the same for both genotype groups ( Figure 1). However, the costs of the tests used during the treatment were double for the 48-week treatment regimen compared to the 24-week regimen (genotypes 1/4/6 vs genotypes 2/3) ( Table 2 and Figure 1).

Figure 1. The cost of the medical tests associated with treating genotypes 2/3 (24-week treatment regimen) and genotypes 1/4/6 (48-week treatment regimen).

A summary of the recommended medical tests at different stages of hepatitis C virus treatment is shown in Figure 1. The unit costs of the different types of tests are shown in Table 3. Costs are in 2017 prices.

Discussion

Treating HCV in Vietnam with IFN or Peg-IFN results in significant direct medical costs. These costs are particularly high because the genotypes that are most prevalent (1 and 6) require a prolonged (48 weeks) duration of therapy ²⁸. The drug-related costs contributed the most (54–89%) to the total treatment cost ( Table 1). The drug-related costs for treating HCV genotypes 1/6 were approximately US$25 per week for IFN plus ribavirin and between US$68-108 per week for Peg-IFN plus ribavirin. The costs of the medical tests and monitoring also contributed notably to the total treatment cost (10–44%) and were related to the duration of the treatment.

The results relate to the cost of a treatment regimen and not the cost per patient cured. Although the cost of treatment with IFN is cheaper, it is less likely to cure the patient successfully and the costs associated with treatment failure can be significant. Consequently, IFN treatment is rarely used in Vietnam.

Studies in neighboring countries have reported the cost of Peg-IFN using the same doses. One study from Thailand reported that the Peg IFN-2a/2b and ribavirin for treating HCV genotypes 1/6 cost US$90 per week (2013 prices) ²⁹, which is similar to our finding. Another study reported that in China ³⁰, the Peg IFN-2a and ribavirin for treating HCV genotype 1, cost US$174 per week (2016 prices), almost double our estimate.

Payment from the national health insurance programme

In Vietnam, the proportion of the population covered by the national health insurance programme as of December 2016 is estimated to be 81.7% ³¹. However, even the patients covered by the national health insurance programme incur significant out-of-pocket costs ( Box 1) for HCV treatment (the co-payment process is outlined in Figure 2). For example, with the maximum level of insurance cover, patients still have to pay 50% of the cost of the IFN drugs and 70% of the cost of the Peg-IFN drugs ( Figure 2) ³². Furthermore, if patients attend the HTD without a formal referral from their primary health care facility, they have to pay for the full cost of the treatment (as though uninsured) ( Figure 2 and Figure 3) ³³.

Figure 2. Summary of the co-payment mechanism for interferon (IFN) and pegylated (Peg)-IFN.

Information adapted from 32, 33.

Figure 3. The possible co-payments required by patients for the hepatitis C virus (HCV) drugs.

The values summarize the different potential co-payment rates by patients for the HCV drugs (based on Figure 3). Note, these only pertain to the drugs and not the other resources/services.

Catastrophic health expenditures

In Vietnam, the per capita average income in 2016 was US$136 per month ³⁴. In comparison, the estimated costs of HCV treatment with IFN or Pre-IFN ranged between US$200 and US$480 per month ( Table 3). Given that even patients receiving the maximum level of support from the national health insurance programme incur substantial out-of-pocket payments for these treatments ( Figure 3), it is likely that many patients will be unable to afford HCV treatment. This may lead to what is known as “catastrophic health expenditures” ( Box 1), when medical spending of a household reaches a point such that the household has to cut down on necessities (such as food, clothing, and their children's education) ³⁵. A variety of different thresholds are used to define this, such as 25% of total household expenditure/income or 40% of a household’s non-subsistence expenditure ^{35, 36}. Regardless of the exact threshold used, our results indicate that HCV treatment is causing catastrophic health expenditures in Vietnam. The importance of reducing such financial barriers is recognised in the Sustainable Development Goals ³⁷.

The move towards using DAAs

DAAs were initially very expensive, thereby restricting their use to high-income countries ³⁸. However, the emergence of low-cost generic versions of the drugs, has led to steep price reductions. A recent study ³⁹ suggested that widespread access to combinations of HCV DAAs is feasible, with potential target prices of US$100–$250 per person for a standard 12-week treatment course, significantly cheaper than the longer treatment regimens with IFN and Peg-IFN. The new DAA drugs are ‘pangenotypic’, meaning they are similarly efficacious for different genotypes, removing the need for expensive genotype testing in specialist labs or prolongation of therapy for the predominant strains in Vietnam. Because DAA treatment regimens are shorter, with fewer side effects, there will also be cost savings associated with the medical monitoring required during treatment compared to interferon-based therapy ( Figure 1).

This cost data related to interferon-based therapy indicates the potential savings of generic DAAs and further supports the economic case to increase access to generic DAAs in low- and middle-income countries.

The most recent Vietnamese MoH treatment guidelines (released in 2016) recommend DAAs as first-line therapy ⁴⁰. However, the costs of DAA drugs only became subsidised by the national health insurance programme in June 2019.

Limitations

Our study has several limitations, for example, we focused on quantifying only the direct medical costs of HCV treatment. This includes the costs covered by the insurance system and the patient’s co-payment ( Box 1). However, the direct non-medical costs (such as the patient's travel costs) and the patient’s productivity costs (indirect costs) were not quantified. The total cost of HCV treatment under the societal perspective ( Box 1) would therefore be even higher. It was also not possible to capture the costs associated with the specific side-effects of IFN/Peg-IFN treatment. In our study, we focused only on patients with HCV infection; in practice, the prevalence of co-infections with other hepatotrophic viruses or HIV can be high, and this is likely to influence the treatment costs.

Our analysis was performed in the context of the HTD in Ho Chi Minh City, which is a large hospital specialising in infectious disease. Whilst it is possible that there may be some minor variations in costs in other provinces in Vietnam, as the costs of both healthcare services and drugs are regulated centrally by the national health insurance department and MoH, our cost estimates are likely to be robust. Although the precise results and cost estimates of our study are not directly generalizable to other countries, they are consistent with reports from neighbouring countries, such as Thailand ²⁹. It is important that further HCV treatment costing studies are conducted in other low- and middle-income countries, particularly relating to the use of DAAs.

The cost estimates were predominantly based on the price lists from HTD relating to 2017. However, it is possible that the costs, particularly those relating to the drugs, will vary over time.

We have focused on conducting a costing study, hence further evaluation regarding the cost-effectiveness of the different treatments is required.

Conclusion

A deeper understanding of the costs of the different treatment options is vital for supporting HCV policy decisions. Currently, there are very few costing studies and economic evaluations of HCV treatment in low- and middle-income countries.

We found that treating HCV with IFN or Peg-IFN results in significant direct medical costs. We estimated that a 48-week Peg-IFN treatment regimen costs between US$3,956–5,887 in Vietnam. The majority of this figure relates to the cost of the drugs.

Although the role of interferon-based therapy is diminishing, this cost data is needed for economic evaluation of new DAA treatment strategies. Many countries are struggling to scale up access to DAAs. This cost data highlights the potential cost savings of using DAAs further supports the importance of increased access to generic DAAs in low- and middle-income countries.

Of concern, we found that even patients receiving the maximum level of support from the national health insurance programme incur substantial out-of-pocket costs for their HCV treatment ( Figure 3). Consequently, many patients will not be able to afford the IFN or Peg-IFN treatments, leading to “catastrophic health expenditures” ( Box 1). This raises important issues regarding the health insurance payment mechanism for HCV patients. Once newer interferon-free regimens are included in the government’s insurance coverage, out-of-pocket expenses for patients could be reduced, but details of how this will be managed are not yet available. Crucially, minimising costs to patients will be an important part of reaching the ambitious 2030 treatment targets ¹².

Data availability

All data underlying the results are available as part of the article and no additional source data are required.

Funding Statement

The MRC GCRF (MR/P025064/1), a Wellcome Trust Collaborative Award (206296) and the Wellcome Trust core grant (106680) supported this work.

The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.