Edited by: Pietro Paolo Michele Iannetta, James Hutton Institute, United Kingdom
Reviewed by: Youssef Aboussaleh, Ibn Tofail University, Morocco; Manisha Choudhury, National Institution for Transforming India (NITI) Aayog, IPE Global Limited, India; Maria Ewa Rembialkowska, Warsaw University of Life Sciences, Poland
This article was submitted to Nutrition and Environmental Sustainability, a section of the journal Frontiers in Sustainable Food Systems
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
It is estimated that more than two billion people suffer from ‘hidden hunger’ (micronutrient malnutrition) globally, with nearly half living in India. Despite being highlighted as one the most cost-effective investments for human development, progress on addressing micronutrient deficiencies (MiND) has been slowing. The severe social, health, and economic costs of MiND in India should make it a top priority for domestic governance and international donors alike. This study, for the first time, maps food system pathways from crop production through to household-level food availability, for a range of key vitamins, minerals, and amino acids. Results suggest widespread (>80% total Indian population) risk of deficiencies in calcium, vitamin A, B12, folate, in addition to lysine limitation, with more localized deficiencies (<25% population) in iron, zinc, and vitamin B6. These deficiencies are the result of a combination of a monotonous cereal-dominated diet lacking in diversity, and overall insufficient food intake. This approach also allowed for “MiND by micronutrient” scenario analysis to 2030, to identify potential intervention points in the food system and the capacity of these interventions to address deficiency. Scenario analysis to 2030 and 2050 indicates that, although increased availability of animal-based products, reduction of supply chain losses, and close to maximum (90%) attainable yields could make some contribution to addressing Indian MiND, additional intervention strategies will be essential. Recommendations for intervention in the short (urgent), near-term (2030), and long-term (2050) have been formulated based on this analysis.
It is estimated over two billion people—more than one-in-three—suffer from micronutrient deficiencies globally (FAO,
Nearly half of the world's micronutrient deficient population live in India (USAID OMNI,
Such deficiencies during pregnancy and in childhood years lead to a range of severe implications including increased mortality, morbidity, physical, and mental defects. Coupled with prevalence of energy-protein malnutrition, India has one of the highest rates of childhood stunting and wasting in the world, occurring in approximately one-third of all children (FAO,
Although progress has been made in addressing MiND in recent years, improvements in South Asia have been too slow to meet Millennium Development Goal (MDG) targets. If current trends continue, SDG2 (achieving Zero Hunger) will also be missed by 2030 (United Nations,
There are a number of interacting social, economic and agricultural drivers of India's micronutrient deficiencies, including relative food pricing of energy-dense vs. micronutrient-rich produce; farmer income effects; and inequitable gender access (Headey et al.,
With a projected population increase to 1.6 billion by 2050 (United Nations,
An efficient and complete food system is necessary to deliver all basic nutritional requirements. Although increasing agricultural production will be a core component in addressing malnutrition in India, alone it may be insufficient to provide adequate micronutrient supply. Recommendations on the capacity of agricultural and alternative strategies for addressing such “hidden” malnutrition might be improved through a holistic analysis of micronutrient production, pathways through the food value chain, and the resultant availability of micronutrients at the household level.
As such, this analysis has attempted to map the flow and pathways of key micronutrients, from crop production to residual food availability within the Indian food system, in order to assess the current and estimate future risk of MiND across the population. Such analysis also aims to explore the capacity to address identified deficiency risks through broad-based food system strategies.
The Indian food system was mapped from crop production through to residual food availability using FAO Food Balance Sheets (FBS) from its FAOstats databases (FAO,
FBS provide mass quantities across the following stages of the supply chain: crop production, exports, imports, stock variation, resown produce, animal feed, other non-food uses, and food supplied (as kg per capita per year). Data on all key food items and commodities across all food groups (cereals; roots and tubers; oilseeds and pulses; fruit and vegetables; fish and seafood; and meat and dairy) are included within these balances. While there are uncertainties in FAO data (see Supplementary Information for further discussion on FAO data limitations), FBS provide the only complete dataset available for full commodity chain analysis. Therefore, while not perfect, they provide an invaluable high-level outlook of relative contribution of each stage in the food production and distribution system.
In order to calculate the total nutritional value at each supply chain stage, commodity mass quantities (e.g., tons of wheat) were multiplied by micronutrient contents from the South Asian FAO INFOODS composition and USDA nutrient databases (FAO,
This study attempts to quantify the average supply and availability of micronutrients through the commodity chain—micronutrients can additionally be lost through processes such as cooking. These latter losses are difficult to quantify and, as such, in the present study, these results are considered to be an upper estimate of micronutrient availability at the point of consumption. It should also be noted that this study assesses only micronutrient availability through dietary food intake; in particular demographics, micronutrients may be supplied in the form of additional interventions such as supplementation or fortification.
In this analysis, the key vitamins and minerals necessary for human health were assessed, including iron, zinc, calcium, vitamins A, B6, B12, and folate. The concentration of iodine in food items is highly variable, and strongly dependent on soil properties (Miller and Welch,
Protein quality is a key concern for India in particular as a result of its largely grain-based diet (Ritchie et al.,
FBS do not provide food loss and waste figures by stage in the supply chain. Food loss figures have instead been estimated based on regional percentages provided in separate FAO literature (FAO,
For consistency, and to provide a better understanding of the food system down to the individual level, all metrics have been normalized to average per person per day (pppd) availability using UN population figures and prospects data (by dividing total nutrient availability at each level by national population figures) (United Nations,
The number of individuals at risk of deficiency for micronutrients and amino acids were quantified using the EAR cut-point method, which is widely applied as the most appropriate for evaluation of micronutrient deficiency (Institute of Medicine,
To assess whether India's micronutrient deficiency risks would decrease through time as a result of expected increases in meat and dairy intake, and continued crop yield improvements under business-as-usual (BAU) policy support, initial analysis (for 2011) was first re-assessed to estimate potential levels of MiND in 2050.
It is projected that, through economic growth and shifts in dietary preferences, meat and dairy demand in India will continue to increase through to 2050. It was therefore assumed that per capita demand in 2050 is in line with FAO projections: this represents an increase in meat from 3.1 kg per person per year (2007) to 18.3 kg in 2050, and an increase in milk and dairy from 67 to 110 kg per person per year (Alexandratos and Bruinsma,
Crop yield improvements were derived based on closure of current farm yields (FY) to reported attainable yields (AY). FY is defined as the average on-farm yield achieved by farmers within a given region, and AY is defined as the economically attainable (optimal) yield which could be achieved if best practices in water and pest management, fertilizer application and technologies are utilized in non-nutrient limiting conditions. Crop yield increases were therefore derived assuming closure of this yield gap to 90% of AY based on published Indian crop-specific figures (Mueller et al.,
Potential climate change impacts on food security and nutrition are numerous, including impacts on yields of staple crops, changes in distribution of pests and diseases, micronutrient density within staple crops and exacerbation of food waste and supply chain challenges (Chakraborty and Newton,
To best demonstrate the food production potential of current agricultural support mechanisms, such as governmental policy and subsidy (which largely determine crop choices), the relative allocation of crop production was assumed constant. It was also assumed that production increases were achieved through agricultural intensification alone; this assumption was based on FAOstats data which has shown no increase in agricultural land area over the past decade, indicating a stagnation in agricultural extensification. To correct for 2050 population estimates, all metrics were re-normalized to “per person per day” (pppd) based on medium fertility scenario projections from the UN prospects (United Nations,
If India is to meet the SDG2 targets of ending malnutrition (i.e., by 2030), these strategies will likely have to be accelerated. To assess the impact of strategic acceleration of food production and waste interventions to meet SDG2 in the context of micronutrient supply, four hypothetical scenarios were assessed. As for the 2050 scenario, all metrics were re-normalized to “per person per day” (pppd) in 2030 based on medium fertility scenario projections from the UN prospects (United Nations,
The following scenarios were assessed:
Following full pathway analysis from crop production through to supply at the individual level, average per capita availability in India in 2011 is shown in Figure
Average per capita availability of key micronutrients and essential amino acids assuming equitable distribution in India (2011). Average per capita availability of essential vitamins and amino acids in India assuming an equitable distribution across the population, measured in 2011. These are measured vs. the national weighted estimated average requirement (EAR). Also noted is the percentage and absolute number of the population at risk of deficiency based on distribution curves. A colored “traffic light” system has been employed, where red indicates a risk of deficiency in >50% of the population; orange for 25–50%; amber for <25%; and green for <5% risk of deficiency.
These results indicate severe deficiency risks across all key micronutrients for the Indian population in 2011. These deficiencies broadly fall into two categories (which are important to differentiate for more effective intervention strategies): “nationwide deficiencies” where the majority of the population are at risk; and “targeted deficiencies” where a smaller and more specific demographic of the population are at risk of falling below requirements.
In this analysis, iron, zinc, and vitamin B6 could be considered to be targeted deficiencies with 41, 25, and 6% of the population potentially falling below EARs, respectively. Estimates of zinc prevalence in 2011 are similar to results published by Wessells and Brown (
Although iron and B-vitamin deficiencies do not result in anemia in all cases, they can act as an important precursor (Lynch,
The other key micronutrients assessed - calcium, vitamin A, B12 and folate – all indicate a widespread risk of deficiency, with 94, 89, 89, and 81% of the population deemed at risk, respectively. An overall lack of dietary diversity, particularly with respect to fruits, vegetables, pulses and animal-based products is likely to be responsible for nationwide risk of MiND with respect to these micronutrients (von Grebmer et al.,
Amino acid availability estimates indicate that lysine, and less notably leucine supply, is limited in the average Indian diet, falling below EAR values. This strongly supports previous studies which have highlighted lysine as a major concern for protein quality, especially in low-meat diets (FAO,
Analysis of how micronutrient pathways evolve from crop production through to household-level availability plays an important role in understanding potential intervention points. There is significant variability in overall pathway patterns between the various micronutrients analyzed in this study. Vitamins and minerals concentrated in highly perishable foods, such as fruits, vegetables and animal-based products, show proportionately higher supply chain losses vs. macronutrients (FAO,
Distribution of Vitamin A, zinc and B12 have been used here to demonstrate the contrast in pathways for nutrients concentrated in fruit and vegetables, cereal-based, and animal-based produce. For example, almost 70% of plant-based vitamin A is lost between crop production and food availability (Figure
Production and losses in the Indian food system from “field to fork” in 2011. Food pathways in
Full results (including availability, requirement, and risk of deficiency) in the case of a BAU agricultural policy and expected (FAO) meat intake to 2050 are provided in Figure
Average per capita availability of key micronutrients and essential amino acids assuming equitable distribution in 2050. Average per capita availability of essential vitamins and amino acids in India assuming an equitable distribution across the population in 2050 projections based on business-as-usual agricultural policies. These are measured vs. the national weighted estimated average requirement (EAR). Also noted is the percentage and absolute number of the population at risk of deficiency based on distribution curves. A colored “traffic light” system has been employed, where red indicates a risk of deficiency in >50% of the population; amber for 25–50%; yellow for <25%; and green for <5% risk of deficiency.
In almost all micronutrients and amino acids, there is a reduction in the percentage of the population at risk of deficiency by 2050 compared to current (2011) levels. This improvement in average availability is sufficient to progress vitamin B12 from a red (>50% of population at risk) to an amber (25–50% of population at risk) rating, and an elimination of leucine limitation in the average diet. However, a risk of severe deficiencies in several micronutrients remains, and lysine continues to be limiting.
It is worth noting that, despite significant reductions in the percentage of the population at risk of deficiency by 2050, the absolute number of individuals at risk increases in most cases as a result of Indian's growing population.
Results of the four scenarios mapped through to 2030 are shown in Figure
Average per capita availability of key micronutrients and essential amino acids assuming equitable distribution under 2030 scenarios. Average per capita availability of essential vitamins and amino acids in India assuming an equitable distribution across the population in 2030 based on meat, waste and yield scenarios. These are measured vs. the national weighted estimated average requirement (EAR). Also noted is the percentage and absolute number of the population at risk of deficiency based on distribution curves. A colored “traffic light” system has been employed, where red indicates a risk of deficiency in >50% of the population; amber for 25–50%; yellow for <25%; and green for <5% risk of deficiency.
Analysis of amino acid limitation highlights that, in all scenarios where expected 2050 meat and dairy intake is accelerated to 2030, lysine availability is no longer considered limiting in the average Indian diet. Note that many individuals will still be consuming less than the average meat intake—for these individuals, lysine, and possibly leucine, limitation would still continue to affect protein quality. The necessity of increased meat and dairy intake for amino acid provision is emphasized by continued lysine and leucine limitation in scenario 2, where supply chain losses are reduced but per capita consumption of animal-based products is assumed to remain at current (2011) levels.
As results from scenario 2 show, a large reduction in supply chain losses is significant in improving availability of vitamin A. This would be expected given that loss of perishable commodities, such as fruits and vegetables, resulted in significant losses of vitamin A from the supply chain. Folate, another vitamin richly concentrated in fruits and vegetables, also showed significant improvements in availability (albeit insufficient to reduce risk of MiND below 50%). Improved supply chain management alone, without increased crop yields and meat intake, would naturally result in an increase in deficiency risks for the remaining vitamins and minerals as a result of a growing population.
The key contribution of an accelerated increase in meat and dairy intake (scenario 1) is an increase in vitamin B12 consumption, as meat and seafood are the only natural dietary source of B12. Vitamin A intake also improves with increased meat and dairy consumption, although this is less effective than a reduction in supply chain losses.
A combination of increased meat and dairy, significant increases in crop yields, and reduced supply chain losses (scenario 4) would result in the largest reductions in MiND risk. Incidence of all deficiency risks, with the exception of calcium, would fall below 50% of the population. Risk of deficiency in iron, zinc and vitamin-B6 would see significant reductions—all below 10% of the population. However, risks of MiND in calcium, folate, and vitamins A and B12 would remain severe.
This study's results indicate that the current risk of “hidden hunger” in India is severe. This has been previously acknowledged within the literature (Klaus von Grebmer et al.,
Analysis of BAU pathways to 2050, and accelerated intervention strategies to 2030, highlight that, while increased meat and dairy intake, increased crop production and a reduction in supply chain losses have the potential to reduce the prevalence of MiND, they will be insufficient alone—even in the most optimistic scenarios—to meet the target of SDG2 by the target date of 2030, or even 2050.
It's important to note the scale of the challenge India would face in accelerating these three broad-based strategies to 2030 as envisaged here. The potential contribution and challenges of each of these options are described below.
Animal-based products are described as “complete proteins,” having adequate proportions of all essential amino acids (meaning none are considered to be “limiting”). In addition to being a key source of high-quality protein, meat is rich in iron, zinc, and B-vitamins; dairy products form a key source of calcium, B12, vitamin A and folate (Rivera et al.,
There is significant agreement that moderate consumption of animal-based produce is particularly important for children, leading to improved growth outcomes, including improved cognition and motor performance (Dror and Allen,
Increased meat consumption has historically been a direct reflection of economic growth (Alexandratos and Bruinsma,
Pulses and legumes may offer a significantly more sustainable alternative protein and micronutrient source (with the exception of vitamin B12; Vecchio et al.,
Supply chain inefficiencies and losses have received significant attention in their contribution to malnutrition (FAO,
The majority of developed countries have planned food processing infrastructure, which has effectively reduced the amount of upstream food loss (although this has transitioned to higher wastage at the consumer level; FAO,
Investment in improved management systems to prevent losses can reap multiple benefits: it improves the nutritional value of foods and subsequently contributes to reducing micronutrient deficiencies; it can allow farmers a higher income through a larger sellable harvest; and it reduces the resource inputs [water, energy, fertilizer, and resultant greenhouse gas (GHG) emissions] for a given utilizable output. The benefits of investment in food supply chain management can therefore be very significant, and reaped by a range of beneficiaries.
Results indicate that the micronutrients with the greatest supply chain losses—vitamin A, folate, and calcium—are associated with widespread risks of deficiency (across the majority of the population in India). This signals the need for a mass intervention strategy with nation-wide coverage. India's demographic distribution currently poses important challenges to developing a country-wide food network that fully addresses MiND. Such infrastructure is often most effective through centralized distribution centers—thereby most-suited to urban populations, and rural regions with sufficient connectivity (Miller and Welch,
This study assessed the impact of closure of current yield gaps to 90% of attainable yields (AY) by 2030 (scenarios 3 and 4) on MiND risks. To achieve this high level of production, India would have to significantly improve on its historical trend of staple crop yield enhancement through to 2030. For example, wheat yields in India are growing at ~0.9% per annum (non-compounding) from 2009 levels and have shown roughly linear growth at this rate over the last decade (Fischer et al.,
Resource constraints in terms of soil fertility (Bhandari et al.,
India's challenge of maintaining balance between macro- (calories, total protein, and fat) and micronutrient (mal)nutrition is difficult to address. India's agricultural policies are currently still oriented toward achieving self-sufficiency in calories and protein (von Grebmer et al.,
This analysis suggests that agricultural policy orientation and land allocation toward production of staple crops may have resulted in a domestic crop composition which is insufficient to also address micronutrient needs. Crop and dietary diversification may offer one option. However, the re-allocation of land used for staple crop production toward more micronutrient-dense commodities will, in most cases, result in reduced total caloric production. This suggests an important conclusion, supported by the results from the scenarios considered here: India's domestic agriculture will be insufficient to address both macro- and micronutrient deficiencies simultaneously.
As such, food imports could play an important role in bridging this gap. However, food imports can have a significant impact on domestic prices (Anand et al.,
The types of commodities essential in reducing MiND vary by micronutrient (key dietary sources of each micronutrient are detailed in Supplementary Table
While the broad-based strategies discussed here could be integral to addressing MiND in India, policies will need to combine these strategies with additional targeted interventions (such as food fortification, biofortification and dietary supplementation). These targeted interventions are detailed by supply chain stage, description and estimated cost in Table
Key targeted intervention options for addressing micronutrient deficiencies in India.
Food fortification | Food processing level | The process of intentionally adding an essential micronutrient to a food, to improve its nutritional quality and provide a public health benefit with minimal risk to health | US$0.05pppa for salt iodisation US$0.12pppa other fortification |
Biofortification | Crop production/field level | The practice of increasing the bioavailable concentration of essential micronutrients in a harvested crop through genetic selection or agronomic intervention | US$1,600,000 per year (national total) for rice in India |
Supplementation | Household level | Concentrated solutions of a particular micronutrient to offer nutritional enhancement to an individual's diet (typically ingested orally as in tablet or powder-form) | US$1-1.20pppa in South Asia |
Results presented in this study indicate an important distinction in deficiency risk between micronutrients: iron, zinc, and vitamin B6 deficiency is likely to be most prevalent in a particular subsection of the population— so targeted interventions which reach the affected demographics (primarily children, pregnant, and lactating women) are therefore necessary. In contrast, inadequate intake of calcium, vitamin A, B12, folate, and iodine are widespread—hence strategies addressing these deficiencies must be implemented across the entire population.
The selection of intervention strategy is context-dependent and determined by several key factors: the specific micronutrient being addressed; the prevalence of deficiency within a given population (i.e., widespread or demographic-specific); and the infrastructural, social and economic circumstances of the country or region in question (Miller and Welch,
The potential role of different intervention measures based on the results of this study, and the Indian context, are discussed below. The feasibility of each strategy for addressing MiND by micronutrient is summarized in Table
Suitability of the various food-based and targeted intervention options in addressing Indian deficiency, by micronutrient.
Iron | x | x | x | x | x | |
Calcium | x | x | x | x | x | |
Zinc | x | x | x | x | ||
Vitamin A | x | x | x | x | x | |
Vitamin B6 | x | x | x | |||
Vitamin B12 | x | x | x | |||
Folate | x | x | x | x | ||
Iodine | x | |||||
Lysine | x | x | x |
Food processing not only allows for a reduction of supply chain losses, but also provides the infrastructure necessary to facilitate food fortification. Food fortification is implemented at the processing stage, and involves the addition or enhancement of one or more nutrients to a food product. Several types of fortification programmes exist, covering mass, targeted, voluntary, and mandatory fortification (Allen et al.,
Mandatory fortification applies in the case where the government makes it a regulatory requirement to fortify a given food product (Allen et al.,
Mass fortification involves the addition of micronutrients to particular food groups or products which are widely consumed across a given population, such as wheat or rice in India. This type of programme is used in addressing nutrient deficiencies which are prevalent across a large proportion of the population. In the case of India, this would include calcium, vitamin A, B12, folate, and lysine. However, this coverage could also be extended to a wider range of micronutrients, especially those such as iron and zinc where deficiency is still highly prevalent, albeit within smaller demographics. The major barrier to mass fortification is India's current lack of centralized food processing and distribution networks; these form a fundamental pre-requisite for effective mass fortification programmes. As with biofortification (described below), the financial hurdle to fortification is the capital cost involved in development of appropriate infrastructure and networks (Miller and Welch,
Food fortification strategies should be coupled with processing developments for reduction of supply chain losses—it is recommended that this forms a near-term (next 5 years) priority, with acknowledgment that coverage is likely to be initially limited to urban populations. Connectivity and wider infrastructure networks for broader coverage should continue to be a focus over longer timescales.
Biofortification occurs at the earliest stage of the food system. It is a comparably newer strategy, involving the innovative use of plant breeding to increase micronutrient concentrations in staple crops(Bouis,
Following the development and distribution of biofortified crop varieties, the farmer should ideally be able to sow and harvest the crop using traditional approaches (i.e., the farmer's only change would be in adopting the new seed varieties) and incur no change in relative costs. Biofortification research and development is still in its relative infancy, with efforts focused across countries in the Global South (Saltzman et al.,
Crops targeted for biofortification should be staple crops commonly produced and consumed by the local population—in India, this is likely to be wheat, rice, pearl millet, and sweet potato. To date, effective biofortification of crops with iron, zinc ,and vitamin-A has been proven, with distribution via the HarvestPlus programme (
The HarvestPlus programme predicts that it could take more than a decade before biofortified crops are widely distributed and utilized in target countries (Miller and Welch,
As with food fortification, investment is largely focused at the capital stage. Limited evidence makes it challenging to complete a total cost-benefit analysis. However, it is estimated that adaptive breeding (capital) costs for biofortification of total rice production in India would be ~US$1,600,000 per year (Meenakshi et al.,
Food processing and biofortification are complementary strategies to address MiND over near- to long-term timescales (>5–10 years). However, the social, health and economic costs of malnutrition in India are on-going, making urgent interventions—such as provision of dietary supplements—necessary to bridge this period. Dietary supplementation is most commonly delivered in tablet or powder-form.
The irreversibility and permanence of maternal and childhood malnutrition means that the most common target groups for dietary supplements are children, pregnant, and lactating women (Stoltzfus,
India's large population size and prevalence of MiND makes the investment scale even more challenging. Supplementation can be inexpensive, with annual costs ranging from US$1-1.20 per person in South Asia and high benefit:cost ratios of (17:1) for vitamin-A supplements alone (The Micronutrient Initiative,
Supplementation should therefore form an urgent and short-term (<5 years) cornerstone in addressing MiND, but should be utilized as a bridge toward more efficient and sustainable delivery mechanisms such as fortification, biofortification, and dietary diversification. Thereafter, supplementation should be reserved for vulnerable demographics with significantly higher daily requirements, such as pregnant women—a practice also implemented in developed countries today.
In summary, results of this study have highlighted serious deficiency risks across most essential micronutrients in India. Scenario analysis suggests that current agricultural policies will be wholly insufficient in addressing micronutrient malnutrition— in fact, orientation toward maximizing macronutrients (predominantly calories) may serve to exacerbate this issue. Broad-based interventions will remain an integral component in addressing MiND in India, with scenario analysis indicating significant potential in the reduction of supply chain losses, and increased dietary diversification through meat and dairy intake. However, India faces a significant challenge in simultaneously addressing macro- and micronutrient malnutrition with a growing population (for reference, rates of 50% MiND in 2030 would put more than 800 million at risk of deficiency in India alone). This limits its domestic potential to increase dietary diversification without causing a negative impact on caloric production.
Such a result indicates that India must address its MiND through an enhanced combination of intervention strategies, including dietary diversification of micronutrient-rich produce including fruits and vegetables, pulses, and animal-based products (domestically and through increased imports), food processing and fortification, biofortification, and supplementation. These interventions are best optimized using complementary approaches, geared toward specific demographics and evolving in line with India's changing socioeconomic and infrastructural development. The high benefit:cost ratios of the MiND intervention strategies considered here should make achieving this enhancement an urgent and sustained focus for the Indian government, and for international aid donors and policymakers.
HR conceptualized the research, developed the methodology, and carried out the analysis with inputs from DR and PH. All authors contributed to writing the paper.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The Supplementary Material for this article can be found online at: