Measurement and mapping of maternal health service coverage through a novel composite index: a sub-national level analysis in India

Background Expansion of maternal health service coverage is crucial for the survival and wellbeing of both mother and child. To date, limited literature exists on the measurement of maternal health service coverage at the sub-national level in India. The prime objectives of the study were to comprehensively measure the maternal health service coverage by generating a composite index, map India by categorizing it into low, medium and high zones and examine its incremental changes over time. Methods Utilising a nationally representative time series data of 15 key indicators spread across three domains of antenatal care, intranatal care and postnatal care, we constructed a novel ‘Maternal Health Service Coverage Index’ (MHSI) for 29 states and 5 union territories of India for the base (2017–18) and reference (2019–20) years. Following a rigorous procedure, MHSI scores were generated using both arithmetic mean and geometric mean approaches. We categorized India into low, medium and high maternal health service coverage zones and further generated geospatial maps to examine the extent and transition of maternal health service coverage from base to reference year. Results India registered the highest mean percentage coverage (93.7%) for ‘institutional delivery’ and the lowest for ‘treatment for obstetric complications’ (9.3%) among all the indicators. Depending on the usage of arithmetic mean and geometric mean approaches, the maternal health service coverage index score for India exhibited marginal incremental change (between 0.015—0.019 index points) in the reference year. West zone exhibited an upward transition in the coverage of maternal health service indicators, while none of the zones recorded a downward movement. The states of Mizoram (east zone) and the Union Territory of Puducherry (south zone) showed a downward transition. Union territories of Dadra & Nagar Haveli (west zone) and Chandigarh (north zone), along with the states of Maharashtra (west zone), Assam, as well as Jharkhand (both from the east & north east zone), showed upward transition. Conclusion Overall, maternal health service coverage is increasing across India. Our study offers a novel summary measure to comprehensively quantify the coverage of maternal health services, which can momentously help India identify lagged indicators and low performing regions, thereby warranting the targeted interventions and concentrated programmatic efforts to bolster the maternal health service coverage at the sub-national level. Supplementary Information The online version contains supplementary material available at 10.1186/s12884-022-05080-5.


Background
Reducing the maternal mortality ratio is a pivotal target as part of the 'Sustainable Development Goal 3' , which affirms 'to ensure healthy lives and promote wellbeing for all at all ages' [1]. Maternal mortality is a crucial health indicator largely resulted from causes which are preventable and treatable [2]. Maternal health, which refers to the health status of a woman during the period of pregnancy, childbirth, and postpartum [3], has the potential of affecting significant maternal and child health outcomes, therefore the domain of maternal health assumes a pivotal role from the development perspective of a nation. Health system strengthening in terms of increased quality of care, health infrastructure, finance, and human resources can help bolster the maternal health service coverage, which in turn can have multiple positive implications on maternal, neonatal and child health statuses. The government of India has initiated many schemes and programmes to improve accessibility, affordability and quality of maternal health service coverage across the country [4][5][6]. Reaffirming its commitment towards achieving better maternal health outcomes, the Indian government, under the National Health Mission (NHM), introduced a novel strategic approach for 'Reproductive, Maternal, Newborn, Child and Adolescent Health (RMNCH + A)' [7]. This 'continuum of care' of RMNCH + A refers to an integrated health service accessible to women, neonatal, infants or adolescents over the dimension of time (before pregnancy, during pregnancy, childbirth and postnatal) and place (family and community care, outpatient and outreach services and hospital and health facilities) [8]. Maternal health thus, is one of the key thematic areas under the strategy of RMNCH + A, which aims to promote linkages between diverse health interventions to reduce maternal, neonatal and child mortality rates.
According to a guideline by the Indian Ministry of Health and Family Welfare (MoHFW), based on the time dimension, the health care services for a pregnant woman are classified mainly into three types, namely; antenatal/ prenatal care, intranatal/intrapartum care, and postnatal/ postpartum care [3]. Antenatal care (ANC), which refers to the systemic care and supervision recommended for women during their pregnancy, is largely a platform for preventive and promotive health care services. ANC primarily aims to monitor the progress of fetal growth and wellbeing of the mother and fetus, thereby facilitating the timely management of obstetric complications.
MoHFW guidelines recommend a minimum of four ANC visits in which the ANC registration and first ANC checkup within the first trimester (within 12 weeks) are accorded a high priority [3]. The World Health Organization (WHO) recommends various interventions ranging from dietary counselling, nutrient supplementation, and essential screening tests to improve the utilization and quality of antenatal care [9]. Another important stage of care in the continuum of RMNCH + A is intranatal/ intrapartum care (INC), which refers to systemic care for a pregnant woman during labour and delivery [3]. The time of birth is considered extremely crucial for both mother and the child as the chances of complications are high during this stage [10]. Sufficient literary evidence supports the prevalence of high-risk pregnancy and pre-term birth among the Indian and other south Asian populations [11][12][13][14][15]. Therefore, the delivery of the baby crucially requires an appropriate setting, where lifesaving equipment is available and hygienic conditions are fully maintained. Utilization of public, private or charitable trust/Non-Governmental Organization health facilities during labour and childbirth, referred to as Institutional delivery, is a significant means of timely management of obstetric complications [16]. However, literature reported that a considerable proportion still deliver at home in India attributed to social, cultural, physical and financial factors [17]. Increased skilled birth attendant (SBA) coverage is associated with a reduced probability of maternal deaths [16], especially for those women who prefer to deliver at home in India. The third stage of care is the postnatal/postpartum care (PNC), which involves systemic care of women after delivery, where the first 42 days (6 weeks) are conventionally treated as the postpartum period [3]. MoHFW guideline recommends at least three postnatal visits to the health facility after an uncomplicated normal institutional delivery in India. Evidence indicates that a larger proportion of neonatal deaths occurs in the first week of postnatal, with the highest number in the first three days [18].
To date, limited literature exists on the measurement of maternal health service coverage at the sub-national level in India. Prior research mainly focused on the theme of Reproductive, Maternal, Newborn and Child Health (RMNCH) service coverage and their inequities at national and sub-national levels in India. The majority of them used Coverage Gap Index (CGI) as an outcome measure of the RMNCH service coverage, which is primarily constituted by the domains of reproductive Keywords: Maternal health, Service coverage, Indian states, Composite index, Antenatal care, Intrapartum care, Postnatal care services, maternal and newborn care, immunization and management of illness [19][20][21]. However, there were only two indicators, namely 'skilled birth attendant' and 'antenatal care coverage' under the domain of maternal and newborn care, which makes the measurement of maternal health service coverage narrow and incomplete. To the best of our knowledge, no published study has estimated and examined maternal health service coverage using a composite index. The present study addresses this lacuna by constructing a comprehensive 'Maternal Health Service coverage Index' (MHSI), which includes a cumulative set of relatively large maternal health indicators categorized into three domains of antenatal, intranatal and postnatal care coverage, which is a major highlight of the study. The use of a composite index value would help us estimate the maternal health service coverage for a particular country/ state and compare them. The MHSI, as a summary measure, will reduce the reporting burden and help in expediting the systematic monitoring of maternal health service coverage in large entities like states and countries. Further, the constructed MHSI has been used for the identification of the low, medium and high coverage regions across different time series points in India, a theme which was largely unexplored in the previous studies. This shall help states identify lagged areas, introduce targeted interventions and prioritize the areas (antenatal, intranatal and postnatal care) to be intervened. Therefore, in this backdrop, utilising a nationally representative data, the specific objectives of the study are outlined as; a) to comprehensively measure maternal health service coverage at the sub-national level in India by constructing a composite index using a rigorous methodology; b) to map and categorize India into low, medium and high maternal health coverage zones; c) to examine the sub-national areas (states and union territories of India) which have transitioned from low to higher category of coverage and vice versa across the different time points.

Data source
The study extracted secondary data for three financial years (2017-18 to 2019-20) on various health indicators of maternal health service coverage from the Health Management Information System (HMIS) portal, an initiative under the National Health Mission (NHM) by the Government of India. HMIS is free of cost, publicly available web-based routine monitoring system of the vital Reproductive Maternal Newborn Child Health (RMNCH) data uploaded by the States and Union Territories (UT) in India. HMIS portal is the only data source that provides time-series data on a number of indicators of the Reproductive Maternal Newborn Child Health (RMNCH) domain in India, which is an added advantage as it allows undertaking temporal analysis and over time incremental changes. The HMIS portal that provides annual data on a financial year basis has been reporting detailed data at the sub-national level since 2017. A total of 29 states and 5 UTs have been included in examining maternal health service coverage at the sub-national level in India during the period 2017-18 (base year) and 2019-20 (reference year). The Indian states and union territories under the study have also been classified into four geographical zones, namely north, west, east & north east and south [22]. The raw and coded data has also been uploaded to Mendeley data repository [23] to provide free and unrestricted access to this time series data for facilitating transparency and reproducibility of the results.

Selection and classification of indicators
The key HMIS indicators at the sub-national level are segregated into different domains, namely child health, family planning, maternal health and the 'other' miscellaneous. Since the present study aimed to construct a composite index for maternal health service coverage at the state and UT levels, the individual indicators have been selected from the domain of maternal health based on the opinion of medical experts and the public health professionals along with an extensive review of the literature. The experts mainly considered the relevance, usefulness, importance, and distinctiveness of the indicators in the valuation of the maternal health service coverage. In consensus with the experts, the study finalized 15 individual indicators to compute the Maternal Health Service Coverage Index (MHSI). It is to be noted that because of the difference in population composition and territory of Indian states, it was not appropriate to choose the indicators reported in the absolute values for the construction of maternal health service coverage. Hence, the study considered the relative measure of the selected indicators for the purpose of index construction.
The selected indicators were further classified under 3 domains of Antenatal care (ANC) coverage, Intranatal care (INC) Coverage and Postnatal care (PNC) coverage based on various guidelines by WHO and the Ministry of Health and Family Welfare [3,9,10,18]. Furthermore, the study classified the component indicators of the ANC coverage domain into 3 sub-domains, namely ' ANC registration & checkups (ANC1)' , 'Supplementation & prophylaxis in pregnancy (ANC2)' , and 'Essential screening tests (ANC3)' . It is to be noted that though all the women do not need c-section delivery, we included c-section delivery as a vital indicator under the domain of INC coverage. This was in view of considering c-section delivery as an important maternal health service indicator, which reflects the very availability of maternal health services, and may contribute to the maternal health service coverage. For instance, if a woman is in need of c-section delivery, but there is no availability of specialist or related facility due to certain reason, in such cases indicator of c-section delivery may become a service coverage indicator. Moreover, with increasing marriageable age [24] and the rising trend of the high-risk pregnancy in India [15], coupled with evolving literature on the need for accounting maternal choices in the type of delivery, have increased the demand for c-section delivery in India [25,26], thereby making it an emerging service indicator for the maternal health. Therefore, in the light of changing times and dynamic health needs, this is another reason for including 'c-section delivery' as an indicator in our study and is one of the novelties of the study. A detailed description of the maternal health service coverage indicators has been given in Table 1.

Data processing
Five percent of the obtained data on the selected 15 indicators of maternal health service coverage were randomly chosen and cross-checked/rechecked for the errors in the data entry to make sure that there was no error in the data entry while copying/transferring the data from the HMIS web portal [27]. Followed by that, the data was checked for the missing values with regard to the identified 15 MHSI indicators. Due to the availability of data for 3 financial years only, the standard imputation techniques and extrapolation and interpolation techniques could not be applied to treat the missing values [28]. The states or UTs with missing values for more than one specific year on any indicator have been excluded from the study. As per this criterion, Daman & Diu and Lakshadweep were removed from the study. The states/UTs having missing values only for one particular year were replaced by the available values of the most recent year [29].

Computation of Maternal Health Service-coverage Index (MHSI) Data normalization
The extracted data on the selected Maternal Health Service coverage Index (MHSI) from the HMIS were subjected to normalization based on the 'Minimum-Maximum approach' used by the United Nations Development Programme (UNDP) for the computation of Human Development Index (HDI) values. Each of the selected 15 indicators is positive in nature, where the resultant higher index value shall represent the positive result (higher the value, better the maternal health service coverage). Owing to the "higher the better" nature of the individual values, the said data was normalised using the following formula [30]: where N i is the normalised value of the i th indicator, A i represents the Actual value of the i th indicator, and U i and L i denote the maximum and minimum values (goal posts) of the dimension indicators, respectively. The UNDP, in its reports [31], recommended the use of absolute goal posts over the observed goal posts in the normalization process. The study fixed 100% (aspirational goal) and 0% (natural zero) as the maximum and minimum values (goal posts), respectively, from which the individual indicators were standardised.

Construction of MHSI: different scenarios
Followed by the calculation of normalized individual indicators, the Maternal Health Service Coverage Index (MHSI) involves three detailed steps. The first step was to estimate three sub-indices of the ANC, i.e., ANC1, ANC2, and ANC3, and to estimate separate indices of INC and PNC. Since both domains of INC and PNC coverage have not been segregated further, the study directly computed INC and PNC index from its individual indicators. Followed by step 1, the study computed the ANC index from its three subindices. Finally, in step 3, the study aggregated ANC, INC, and PNC indices to compute a comprehensive cumulative index for 'Maternal Health Service Coverage' .
The study employed four different scenarios using the aforementioned three steps for the computation of MHSI. The different scenarios have been explained in detail in Fig. 1. Further, an additional file 5 clearly explains the formula used in each step of four different scenarios dealt in the study. Scenarios differ by choice of aggregation method, i.e., usage of arithmetic mean approach and geometric mean approach [31], which is primarily based on the procedure followed for the construction of the Human Development Index (HDI) by the United Nations Development Programme (UNDP). In scenario I, all the sub-indices and domain indices of MHSI were aggregated through the arithmetic mean approach. It is a basic and popular method of aggregating the indicators. However, the major disadvantage is that the results are largely affected by the extreme values and are not robust in nature [31]. However, the usage of arithmetic mean is more suitable in those cases where the data does not contain outliers or skewed data and normalization is carried from the observed values of the data distribution [31,32]. In scenarios II and III, we used both arithmetic mean and geometric mean approaches, as detailed in Fig. 1 and Additional file 5. Since these two  The total number of reported deliveries scenarios also incorporate the use of the arithmetic mean at different stages, it is also plagued with similar problems as highlighted above (less robust for policy formulations). In contrast to scenario I, scenario IV aggregated all the sub-indices and domain indices of MHSI through the geometric mean approach. The geometric mean approach has an advantage over the arithmetic mean approach (used in the scenario I) as it is less affected by skewed data and extreme values, thereby integrating the overall balance in the data distribution [31]. It is considered to be a robust and higher-level approach as per the UNDP methodology for constructing a composite index, therefore more suitable for decision making and policy interventions. Microsoft Office Excel 2019 and IBM SPSS version 24 were employed for the construction of the index and computation of related descriptive statistical measures.

Categorization and mapping of Indian states and Union Territories
The index values of the MHSI ranges between 0 and 1, where values close to 1 indicate high maternal service coverage and value close to zero correspond to low coverage. Furthermore, the study categorized Indian states into low, medium, and high categories based on the MHSI score. The study considered the period 2017-2018 as the base year and calculated the 33 rd and 66 th percentile scores of the MHSI values of Indian states. These percentile values were used as the criteria for determining the category Indian state belonged to. The Indian states with an MHSI value below the 33 rd percentile were categorised to the low category, those with above 66 th percentile were categorised to the high category, and the ones having a score between 33 rd and 66 th percentile were assigned to the medium category. The study categorised

Results
Descriptive characteristics of maternal health service coverage indicators Table 2 and Fig. 2 describe the characteristics of selected maternal health service coverage indicators at the regional and national levels. Out of the 15 indicators, institutional delivery showed the highest mean percentage across India and every region. The study observed that, on an average, 93.7% of the total reported delivery delivered at the institution (both public and private) in India. Across India, 'treatment for obstetric complications' was observed to be the lowest (9.3%) among all the indicators.
On the other hand, the north zone registered the lowest mean percentage coverage for immunization against tetanus (73.6%), screening for HIV (37.9%), SBA attended home deliveries (17.3%) and postpartum checkup (54.7%). Among the indicators, none of the zones recorded more than 50% coverage in screening for diabetes, SBA attended home deliveries, c-section deliveries and treatment for obstetric complication. Furthermore, estimates of the coefficient of variation showed indicator of 'institutional delivery' has the lowest and the indicator of 'the screening for diabetes' has the highest variations in India.

Composite Maternal Health Service Coverage Index (MHSI) and its comparison between the base year and reference year
The study computed MHSI using different methodological based scenarios already explained in Fig. 1. The MHSI values based on four different approaches for the base year (2017-18) and reference year (2019-20) have been presented in the Additional file 2. In scenario I, all the sub-indices and domain indices of MHSI were aggregated through the arithmetic mean approach, whereas, in Scenario IV, the geometric mean approach was employed [31]. For conciseness, MHSI values across the base year and reference year were compared only for scenarios I (Table 3) and IV (Table 4). However, in the interest of the large audience, details of MHSI values across the base and reference years for scenarios II and III have also been presented in Additional files 3 and 4. Nevertheless, these tables do not form the part of analytic inferences of the present study.
For scenario I (Table 3) For scenario IV (Table 4), which is based on the geometric mean approach, the MHSI value for India had Table 2 Descriptive characteristics of indicators of maternal health service coverage at regional and national level over a span of three years (2017-18 to 2019-20) and reference year has also been presented through geo-spatial maps in Fig. 3 (Fig. 3A for scenario I and Fig. 3B for scenario IV).

Discussion
Out of the 15 maternal health service indicators, institutional delivery (INC indicator), immunization against tetanus, ANC checkups and IFA supplementation (ANC indicators) recorded a relatively higher mean percentage (more than 70%) coverage at all India level for three years (2017-18 to 2019-20). Maternal health care services like institutional delivery and antenatal care coverage in India observed substantial progress in the past two decades with the implementation of plethora of targeted government driven programmes such as the National Health Mission (NHM), Maternity Benefit programmes  [4,5,33]. Recently launched ' Anemia Mukt Bharat' programme has given prime importance to IFA supplementation for pregnant women, which is a highly probable reason for   recorded relatively low mean percentage (around 30%) coverage across the country. Some state specific studies recorded prevalence of gestational diabetes ranged from 9.89% to 35% [35][36][37][38][39], which is quite alarming. In this regard, in compliance with the recommendations of WHO and the International Association of Diabetes and Pregnancy Study Group (IADPSG), India requires region specific efforts to expand its coverage of screening for diabetes among pregnant women [38]. Though the present study observed lower mean percentage coverage for c-section delivery across India, nevertheless the growth rate of c-section delivery in India has doubled during 2005-06 to 2015-16 [40]. Low socio-economic status, awareness of available schemes, distance to the health facility, absence of transport, and social cultural norms can be identified for preference of home delivery and related maternal complications in India [41][42][43][44].
Both southern and western zones of India have outperformed the other zones on the coverage of a number of maternal health indicators. South zone, particularly has performed sizable well with respect to institutional deliveries, IFA supplementation, ANC checkups and screening for HIV [45][46][47][48]. The states of Kerala, Tamil Nadu, Karnataka and Andhra Pradesh from the south zone are the front runners in the overall 'Sustainable Development Goals' (SDGs) with relatively high SDG index scores, which has remarkably influenced their coverage and receipt of health care services such as institutional care and antenatal care coverage [49]. Consistently high HIV prevalence districts were clustered mainly in the south zone of the country [50], therefore, it is expected to have high percentage coverage in screening for HIV as a preventive measure. Furthermore, most of the states from south zone are known for their remarkable performance in literacy rates, particularly in female literacy [51], which can be a possible reason for increased awareness of HIV screening among south zone population. West zone particularly has recorded a substantial high coverage for immunization against tetanus, ANC registration, calcium supplementation and anemia treatment coverage. Effective planning and management of resources, re-establishment of outreach services, community links, supportive supervision and monitoring are the possible facilitating factors of immunization services including tetanus [52]. State governments of Gujarat, Chhattisgarh and Madhya Pradesh from the west zone have collaborated with WHO for calcium supplementation to pregnant women [53] possibly the reason for high 'calcium supplementation' coverage in the west zone.
North zone recorded the lowest coverage for indicators particularly with respect to screening for HIV and SBA attended home deliveries. Low female literacy and rigid socio-cultural factors of the majority of states from the north zone may have resulted in low SBA attended home deliveries [51]. However, Uttar Pradesh and Rajasthan from the north zone are observed to be the states which require the highest need for 'Prevention of Maternal to Child Transmission' (PMTCT) of HIV [54] and therefore demands high coverage in screening for HIV among pregnant women. East & north east zone has underperformed on a number of indicators, specifically for screening pregnant women for diabetes, c-section delivery and treatment for obstetric complications. The low female literacy rate [51] and the high proportion of tribal population [55] are expected to account for low c-section delivery rates in this zone. Majority of states from the east & north east zone recorded a low prevalence of diabetes in a nationwide study [56], making it a likely cause for low diabetes screening coverage. Inaccessible regional terrain, poor transportation facilities, availability of prompt medical services and behavioural factors (tobacco and alcohol use) are significantly associated with a low treatment coverage for obstetric complications [57].
The methodology adopted in the present study for data normalization and construction of MHSI is primarily based on United Nations Development Programme (UNDP)'s procedure to compute Human Development Index (HDI) [58], which is highly rigorous and standardized. Further, the study generated MHSI scores based on the choice of arithmetic mean and geometric mean approach for aggregating the 15 indicators into one composite index. The geometric mean approach (used in scenario IV) has an advantage over the arithmetic mean approach (used in scenario I) as it is less affected by skewed data and extreme values, thereby integrating the overall balance in the data distribution [31]. However, the usage of arithmetic mean is more suitable in those cases where the data normalization is carried from the observed values of the data distribution [32]. Therefore, comparison of index values across these two different approach-based scenarios are redundant in nature. We clarify that the study elucidates the comparison of MHSI values across two time periods (base year and reference year) within a given scenario and not across the two scenarios. Construction of MHSI for these two particular scenarios, facilitates the identification of those common states/union territories/regions, which have transitioned form a lower to higher category of maternal health service coverage or vice versa. The earlier attempts to measure maternal health service coverage have been limited and narrow in their approach. Composite Coverage Index (CCI), which comprised essential health interventions of RMNCH continuum of care [59], included only two indicators i.e., 'skilled birth attendant' and 'antenatal care coverage' under the domain of maternal and newborn care. In contrast, the present study specifically emphasized on maternal health service coverage rather than considering the consolidated RMNCH coverage as studied in the previous literature [19-21, 59, 60].
Different states/UTs/zones in India have performed variedly as far as MHSI scores are concerned. The UTs of Dadra & Nagar Haveli and Chandigarh showed upward transition of the MHSI category in both scenarios (I and IV). Dadra & Nagar Haveli transitioned from medium to high category and that of Chandigarh showed low to medium category. This might be likely due to the concentrated efforts of respective governments and also partly because these UTs are the front runners or performers in specific SDGs related to poverty, health, quality education and economic growth [49]. The state of Maharashtra and Assam showed an upward transition in both scenarios from medium to high MHSI category, while Jharkhand showed upward transition from low to medium category. Assam and Jharkhand have remarkably improved its maternal health outcomes and have recently been accorded a position of 'achiever' in a health index released by the National Institution for Transforming India (NITI) Aayog, which is the Indian government's apex think tank [61]. While, the state of Maharashtra is a frontrunner across various dimensions of NITI Aayog's health index. On the other hand, the state of Mizoram and the UT of Puducherry displayed downward transition in both scenarios. This might be due to relatively less coverage on a number of indicators like IFA supplementation [62], diabetes screening [56,63], ANC checkups and treatment of obstetric complications. Out of the 4 zones, only the west zone transitioned upwards in the reference year from the medium to high MHSI category. The state of Goa and Maharashtra and the UT of Dadra and Nagar Haveli from the west zone performed substantially well, which gave an upsurge in the MHSI scores of the west zone. None of the zones showed downward transition across two time points for both the methodological scenarios. Region/state/UT specific health interventions are required to be undertaken to target low coverage indicators of maternal health for bolstering their performance to achieve positive incremental changes in maternal health outcomes over time.
The present study faces some limitations. The present study is based on the HMIS data that allows undertaking temporal analysis and over time incremental changes, nevertheless some issues still remain in terms of data quality and therefore the results of the present study should be cautiously used. Consideration of private sector, which reports at a lower rate than the public sector to the health facility data like HMIS, may affect the numerators of the MHSI individual indicators especially for the states with a greater predominance of private sector. Therefore, readers have to be watchful while drawing and generalizing the conclusions. There might be inherent issues with some HMIS derived denominators used to calculate the coverage estimates. The denominators indicates those females who were registered for a maternal services for instance 'Total number of pregnant women registered for ANC' . It excludes that population who have not been registered for the said maternal service or did not seek maternal health care elsewhere. Those women might be marginalized, more vulnerable and lesser health status. This can lead to selection bias depending on proportion of women registered for any service. The external validation of MHSI individual indicators from HMIS data with that of state specific data or NFHS data and alike should be undertaken for better generalizability, which forms the future scope of this study. Besides, due to limited data availability at the sub-national level, the analysis pertains to a short time span of three years (2017-18 to 2019-20), hence the incremental changes in the maternal health service coverage reflect only the preliminary estimates, therefore should not be used to draw inferences for long term policy purposes. More time series data is required to confirm and substantiate the inferences of these short time series-based results. This leaves a lot of scopes for future research to undertake longitudinal analysis concerning the theme of the present study. Though the composite 'Maternal Health service Index' computed in the study involves a set of 15 indicators, which is relatively larger than the previous studies, however; data on only one indicator (postpartum checkup) was available and classified under the domain of postnatal care, which makes this domain under-represented for constructing the index. The government of India has initiated a number of schemes especially for nutritional interventions of pregnant and lactating women [64,65] for antenatal and postnatal care. In this light, besides postpartum checkup, which is a globally recognised indicator [66], data on more indicators like nutritional support, screening for psychological wellbeing after delivery [67] and alike should be disseminated on the government portal for better representation of the PNC domain, which is an important food for thought for a revised version of this novel index in future. Additionally, though the study included 'c-section delivery' as an important indicator under the domain of intranatal care; nevertheless, the states need to be cautious and should take into account the trend/ prevalence of high-risk pregnancy, obstetric complications and evolving concept of maternal choice/ decision while taking a decision on the inclusion of particular indicators like c-section delivery as a coverage indicator as all the women are not in need c-section delivery. Although the study envisioned to examine each and every state as well as union territories, however the two union territories of Daman & Diu and Lakshadweep were excluded from the analysis as they both registered missing values for more than one specific year on a particular maternal health indicator. Nevertheless, the present study has some strong points and strengths to offer. The novel 'Maternal Health Service Coverage Index' envisaged and constructed in the present study is a pioneer attempt to comprehensively quantify the sub-national coverage of maternal health service in India by including within its ambit 15 indicators spread across each of the three domains of antenatal care, intranatal care and postnatal care. No published study has elucidated the measurement of maternal health service coverage in terms of the broad spectrum of antenatal, intranatal and postnatal care coverage till date in India, though these domains hold prime importance and are widely valued by the reputed international and national organizations [3,9,68]. There is wide scope of usage for this unique 'composite index' as it provides a concise and quantified information of complex and multifaceted facts representing maternal health service coverage, which makes it easy to communicate and report for the policy purpose [69] at the sub-national level in the Indian setting. This composite summary measure, will reduce the reporting burden and speed up the monitoring of maternal health service coverage at sub-national level in India.The study has further mapped and categorized India into different zones, thereby providing a snapshot of low, medium and high maternal health service coverage states/union territories and regions. This shall help states identify and prioritize the laggard areas (within antenatal, intranatal and postnatal care) to be intervened. Furthermore, the study by considering a nationally representing data published by the Indian government, has facilitated the comparison of the maternal health service coverage performance at the sub-national level across different time points, which has allowed to identify which states/union territories and regions have transitioned from lower to a higher maternal health service coverage category and vice versa.

Conclusion
Measuring and mapping of maternal health service coverage at the sub-national level can have significant policy implications for a developing country like India, which is making rapid strides in healthcare sector transformation. A composite summary measure to reflect the coverage of maternal health services can momentously help India to identify low coverage maternal health service indicators and recognize lagged zones, thereby warranting the concentrated programmatic and clinical efforts to detect, examine and target specific characteristics of a particular state/UT/region that has a significant effect on maternal health service coverage. The accessibility, availability and utilization of quality driven maternal health services are key for survival and wellbeing of both mother and child. A regular longitudinal and extensive monitoring of the performance of states with regard to essential maternal health services shall induce the country towards achieving the SDG target of improved maternal and child health outcomes.