IV. Improving Accountability: The Critical Need for Better Monitoring and Timely Investigations

While Indian authorities have taken steps to improve healthcare system accountability, existing approaches have not done enough to ensure that they know where the problems are. The key issue here is effective monitoring: using district-level investigations and appropriate monitoring indicators to obtain the data needed for effective interventions to reduce maternal mortality and morbidity. 

Central and state authorities often point to the number of facility-based deliveries as an important measure of progress. While this is a partly useful measure-facility-based deliveries do correlate with reduction in maternal mortality if certain preconditions are met-it is by no means sufficient and officials have relied on it too much.

We believe that more attention must be given to timely district-level investigations into maternal deaths, to use of so-called "UN process indicators" on emergency obstetric care, and to improved reporting of deaths, including through the civil registration system (which records key life cycle events such as births and deaths). This chapter, drawing heavily on Human Rights Watch research in Uttar Pradesh, analyzes shortcomings in these areas.

Poor District-level Monitoring

Monitoring at a level "where there is power to effectuate change"[159] is key to rectifying health system shortcomings. UNICEF has observed that a "national policy requiring specific notification of maternal deaths" would be a powerful tool in reducing maternal mortality.[160] Public health experts and researchers have shown the utility of investigating maternal deaths to get a better understanding of and information about the socio-economic and medical causes of maternal deaths.[161]

Under the NRHM, the Indian central government has recognized that maternal death investigations are a powerful "monitoring tool" at the micro level,[162] reiterating that "reduction of IMR/MMR will also be closely monitored through social audit." The erstwhile RCH-II program now part of the NRHM also requires the collection of information on maternal deaths. In practice, however, such data collection and processes have thus far played little or no role in planning and evaluating maternal healthcare interventions in Uttar Pradesh and many other parts of the country.

The Indian central government and many states have taken concrete steps to improve surveillance of maternal health. Dr. Jorge Caravotta, health specialist from UNICEF India, said,

The Ministry of Health and National Health Systems Resource Center are [working] with UNICEF, UNFPA, WHO to create a maternal surveillance system in the country. Facility and community-based maternal death audits now will be a tangible reality in the field to improve health systems and reduce maternal mortality. The Indian government is taking steps to develop implementation guidelines for this purpose. Ten states have included maternal death audits in the state project implementation plans."[163]

Effective investigations of maternal deaths must be designed to identify systemic issues in healthcare provision which may be causal factors in maternal deaths, particularly among women from marginalized communities.[164] The UN Special Rapporteur on health, after his visit to India in 2007, underlined the utility of such investigations. Urging all states to introduce a system of maternal death audits, he stated that it is of "utmost importance that all the circumstances of maternal deaths be examined in order to find out why the death occurred ... [T]hey can help to identify the structural and systemic failures that are leading to women's preventable deaths."[165]

Barriers to Accurate Reporting of Maternal Deaths

Even though the Uttar Pradesh state government is yet to include maternal death investigations in its state project implementation plans, it has demonstrated its commitment to monitoring adverse pregnancy outcomes through three important initiatives. First-in a pre-NRHM initiative-the government issued a 2004 governmental order requiring maternal death audits, but it was not implemented.[166] Second, in 2007 the state government, through the World Bank funded Uttar Pradesh Health Systems Development Project, undertook a pilot "Maternal Death Notification and Review System."[167] Third, the state government is in the process of launching UNICEF's Maternal and Perinatal Death Enquiry and Response (MAPEDIR) initiative in several districts.[168]

As the Uttar Pradesh government gears up to implement the recently launched the NRHM Health Management Information System (HMIS) and develops initiatives to investigate maternal deaths, it needs to address the gaps in its reporting systems to ensure accurate data collection, particularly on maternal deaths irrespective of place of death-home, en route a health facility, public or private health facility. Health officials should avoid the same pitfalls that plagued the implementation of the 2004 government order. A former senior health official explained that committees to conduct maternal death audits were created and ANMs were supposed to report deaths, "but communities were not involved and no information was given to anyone."[169]

Poor political commitment and priority for such data, a lack of awareness about the objective and importance of such reporting, a lack of adequate training for health workers to implement the maternal death audit system, and underreporting of deaths due to fear of punitive action were cited as reasons that hampered the implementation of the government order.[170] A "lack of a standardized, effective and efficient information system which also includes maternal death tracking for planning, monitoring, and action... is one of the major challenges" for maternal health in Uttar Pradesh.[171] 

The Uttar Pradesh state government should immediately address the problem of gross underreporting of maternal deaths. A pilot study conducted by the Uttar Pradesh Health Systems Development Project in a few blocks across four districts identified roughly 380 deaths within six months.[172] While one area registered 100 percent reporting, most areas were found to have average, below average, or poor reporting of maternal deaths.[173] A nongovernmental organization identified 60-70 maternal deaths in one block of a district (administrative area with 100,000 to 300,000 population) between 2006 and 2008.[174] Human Rights Watch spoke with three of the families who had suffered maternal deaths in 2008.[175] The district medical authorities were not aware of any of these deaths. Instead, when pressed, they provided cursory "unofficial" informationaboutfour or five deaths in the whole district (group of blocks) for 2008; the official record reported "zero" maternal deaths.[176]

When asked about the number of maternal deaths in his district in 2008, another district health official said,

Number of maternal deaths-now that is a million dollar question. None reported. But we have heard of 12 or 13 maternal deaths through word of mouth. There was one in the district hospital ... In those other 11 cases, they are afraid to report deaths, they think it will give them a bad name.[177]

Similarly, in Chitrakoot district, the official report showed that there were no maternal deaths for the past year,[178] but Human Rights Watch documented several maternal deaths.[179]

Rae Bareilly district exhibited a better system of reporting deaths. District health records from April 2008 to January 2009 showed 107 maternal deaths, including two deaths in January 2009.[180] The district health official who showed us the records said that the data suffered from underreporting, and was eager to compare the government figures with those generated by PATH, a nongovernmental organization documenting maternal deaths in the region.[181]  

Unless deaths are reported, they cannot be investigated and acted upon. The recording and reporting of maternal deaths at so-called sub-health and primary health centers is a "service guarantee" under the NRHM.[182] But to translate guarantees on paper into a "culture of reporting, investigating, and acting on maternal deaths," the government should create an enabling environment, dismantling the barriers against reporting and investigating maternal deaths.[183]

Low Priority for Data on Maternal Deaths

Poor "demand from above" for information on maternal deaths, activists and government officials felt, was a key reason for the poor state of reporting and investigation of such deaths at the field-level.[184]  An expert group that conducted a maternal death notification pilot study for the Uttar Pradesh Health Systems Development Project found that "reporting and recording deaths is not a priority for any department,"[185] and that there was "poor initiative from health sector on notification [of maternal deaths] activity."[186]

"There is no data on maternal deaths because no one bothers to collect it," said Dr. Neelam Singh, a Lucknow-based gynecologist and activist experienced in investigating maternal deaths. Health workers on the field responded to what she called "the danda [stick] approach-if the data is demanded from the top, then only the district officers will get into action and put pressure on their juniors to give such data."[187] Many activists and government officials used the polio eradication campaign as a "classic example" illustrative of the top-heaviness of the health system. Cautioning that the coercive element of the polio eradication campaign merits attention and should not be repeated, several health experts and government officials stated that the key lesson from the polio eradication campaign was that it demonstrated that the public health system can be made to work.[188]

The low priority given to data on maternal deaths became evident to Human Rights Watch when senior officials from the directorate of family welfare appeared unaware of their own reporting formats. What directorate officials told us was directly contradicted by workers in the field.

One senior official told us: "This information [maternal deaths] doesn't come to us because we don't get this through the pro forma. We don't have a column for maternal deaths."[189] Describing what he called an error, one health official explained 2008 revisions in their reporting formats. He said,

When we used to have CSSM forms [Child Survival and Safe Motherhood forms], under "Surveillance" we used to have a maternal deaths column. From last year we have given new forms-called routine immunization now- but most of the data collected in this form is also the same-about deliveries also. But the maternal deaths column in this form is missing-I think it got left out by mistake.[190]

Health workers and district-level officials, however, say they are required to provide maternal mortality data to the directorate. One district official showed Human Rights Watch a form titled "Monthly Report to Monitor, Components of M.C.H. Programme."[191] This form contained a column labeled "number of mothers died out of the above," against which district officials are expected to furnish information for the current month, as well as an aggregated total from April of the preceding year. The February 2009 report submitted by one district under this format showed that they had "nil" deaths from April 2008 even though doctors and families shared with Human Rights Watch at least three cases of maternal deaths in this district between April 2008 and February 2009. This indicated that either the district officials were completely unaware of the number of maternal districts in their district or that they were suppressing information about such deaths.[192] ASHAs and ANMs also showed Human Rights Watch their reporting registers that carried a column for recording maternal deaths.[193]

Several health workers reported, however, that they were not asked about maternal deaths during staff meetings. For instance, describing her monthly review meetings, one ASHA said,

[E]very month we [ASHAs] have a meeting in the CHC. All ASHAs are called. At this meeting we discuss nasbandhi [sterilization] and teekakaran [immunization]. Nothing else.[194]

Similarly, explaining how her review meetings were conducted, an ANM said,

[Our] supervisor meets about 20-25 ANMs together. Issues surrounding maternal deaths are not discussed. Last year there was a maternal death in the CHC ... I do not know more details. This case was not discussed in our ANM meeting.[195]

A staff nurse who reported a recent maternal death in her health facility in January 2009 said, "This death case was not shared in any common meeting. We don't have any [such] system."[196]

District health officials stated that they collect information through field-based health workers and generate monthly reports about the number and place of deliveries.[197] The fact that the same district health machinery regularly generates estimates of deliveries and place of delivery, but repeatedly shows near-zero reporting for maternal deaths is indicative of the poor priority and demand for such data.[198]

Several Uttar Pradesh health officials claimed that they struggled to improve their reporting because of their large population. Dr. T. Sundararaman of the National Health System Resource Center, however, felt that population was not a barrier to improving reporting mechanisms. He said that political will and better governance were required to improve data collection.[199]  Similarly, Dr. Vishwajeet Kumar, an expert on infant and maternal mortality said that governance, not population, was the issue. He said, "underreporting is a consequence of un-intentioned methodological limitations and seemingly deliberate administrative opaqueness to maternal deaths."[200] 

Lack of Definitional Clarity and Inadequate Training

Kavita K. developed complications and died in December 2008 about two weeks after her delivery.[201] Health officials were likely oblivious of her death because of several reasons-one of them as simple as poor awareness of the definition of a maternal death.

Accurate reporting of maternal deaths requires definitional clarity. Many field-based health workers stated that they were the primary sources of information about maternal deaths in villages.[202] Even though ANMs are primarily responsible for the care of pregnant women, many health workers reported relying on each other for information regarding pregnant women and maternal deaths.[203] Almost all of them, including ANMs and anganwadi workers with more than 20 years experience, were unfamiliar with the definition of a maternal death.[204]

Identification of a maternal death hinges on two things: its timing and cause. Depending on when a pregnant woman dies-either during pregnancy or within 42 days after termination of pregnancy (postnatal period)-the death is classified as pregnancy-related.[205] All such deaths are then filtered by medical cause, determining whether the pregnancy had any direct or indirect role in the death.

Maternal deaths are most frequent in the postnatal period, that is, within 42 days of termination of pregnancy.[206] But where health workers are not trained to track and report any pregnancy-related death for the entire period within which they are likely to occur, particularly the postnatal period, deaths will go unreported or misclassified. Health workers from three districts we visited gave varying descriptions of the postnatal period, ranging from a few hours after delivery to a month.[207]

Ratna R. and Kishori K., anganwadi workers who are charged with providing nutritional supplements to children and pregnant women in rural areas, said they kept 15 different registers. In one of them, which they described as the "birth and death register," they said that they were supposed to record details of deaths. One said,

In this we note down the name of the person who died, date of the death, age, reasons-we note down if it is a child, but adults also sometimes we note down. If it is a pregnant woman who died then we note it down-we have to report it-any death during delivery or after delivery-within 6 or 8 hours after delivery ... If it is after that then we write the reason-there will be other reasons-fever or something else. Those are not maternal deaths. How can those be maternal deaths? [208]

Vibha V., an experienced ANM, explained that they were supposed to record maternal deaths. "It is all in the same register, but it has different parts," she said, directing Human Rights Watch's attention to that portion of her register that is titled "Maternal Deaths." "See this," she said, showing the empty pages of the portion of her register for maternal deaths, "this is where we are supposed to record it." She had not recorded any deaths including one documented by Human Rights Watch in the area where she worked. When asked what it means to have a maternal death, she said, "It could be the death of the mother when the child is still inside her and she has not delivered, or during delivery." When Human Rights Watch prodded further asking whether such deaths could occur after the delivery, she nodded in affirmation and said, "It could be after delivery also-that is during the navjoth period [within 28 days of delivery]. If it is within this period we say it is maternal death."[209]

Such a lack of definitional awareness prevailed even though the Uttar Pradesh government had distributed reporting registers to ANMs and ASHAs with clearly printed columns indicating the different periods in which a pregnant woman or mother could die. The reporting format contains a column labeled, "When did the pregnant woman die," that is further subdivided into "before delivery," "during delivery," and "within 6 weeks of the delivery."[210] 

Poor translation of textbook definitions into practice is indicative of several things, including  lack of understanding of the importance of such definitions and the importance of using registers to report cases. Insufficiently frequent training and refresher courses, several doctors and government officials felt, influenced how health workers functioned, including their ability to track maternal deaths accurately.[211] Emphasizing the importance of such courses, a senior officer pointed out why health workers are likely to forget the definition of maternal death and its reporting procedures: "In real terms ... maternal death in a village may occur once in one year, or sometimes once in two or three years. By the time the death occurs, she [health worker] will have forgotten everything."[212]

The Uttar Pradesh state government gave training a low priority until 2008. Since then health officials have apparently begun reviving training programs for ANMs.[213] The 2007 government Third Joint Review Mission reported that "after a gap of more than a decade" the Uttar Pradesh government is planning to restart training for ANMs.[214] A former senior health official provided an insight into the challenges he faced reviving ANM training programs:

For most of the past 15 years there has been no recruitment of ANMs. In 2004 a few ANMs were recruited under the reserved category. After that again under NRHM some ANMs have been recruited on contract-basis. Because of this the ANM training centers were lying in complete disuse for more than a decade. Many of them were being used as offices or storage. At least 26 of the 40 ANM centers were restarted recently to conduct training sessions for ANMs. Without regular training there can be no skill building.[215]

ASHAs stated that the state government had organized two training programs since January 2007,[216] but one of them said that the time-lag in distributing training manuals and registers adversely affected her data collection in the field:

We had one training in January 2007 and another in March 2008. They give us the training first and then after many months they give us the books [training manuals and registers]. I got the books in September 2008. The same is the case with the survey [of pregnant women] register. They give us the survey register midway through the year so they will not get a proper survey.[217]

Lack of Continuous Care through the Antenatal and Postnatal Periods

Unless there is sufficient emphasis on continuous care, including care during the antenatal and postnatal period, deaths occurring in these periods cannot be prevented, much less reported. Given that ASHAs and ANMs keep all primary information regarding pregnant women, if they are not in regular contact with pregnant women in their areas, then the chances of their providing assistance or documenting deaths accurately are significantly reduced.[218]

For instance, Human Rights Watch documented two cases where women developed complications during the antenatal period. In both cases health workers were not in regular contact with the families during the antenatal period as required by NRHM norms and the women ultimately died. In one case in October 2008, Meena M.'s daughter-in-law, Aditi A., started bleeding in the third month of her pregnancy and was referred from one private health facility to another before her family finally managed to have her admitted at a private health facility in Lucknow city where she died.[219]

In another case in June 2008, Munira M. from Chitrakoot district started bleeding in the eighth month of pregnancy, got referred from one private facility to another, and finally died in a government hospital in Allahabad district.[220] When asked whether families were aware of any initiative to report or register the death, they stated that no one had come to find out details about the death, suggesting that it went unreported. Relevant district medical officers stated that there had been no maternal deaths in 2008.[221]

Similarly, Human Rights Watch documented four maternal deaths in the postnatal period[222] in which families reported that no one had come to note down the details of the death, suggesting that they went unreported.[223] Human Rights Watch can confirm that at least two of the four deaths went unreported.[224] 

Death of Aditi A., Unnao district, October 2008 (Story as narrated by her mother-in-law and neighbor) Aditi A., about age 20, was pregnant for the first time. The ASHA and ANM used to come to her village for "iron golis [IFA tablets] and teekakaran [immunization]." According to Niharika N., Aditi's mother-in-law, the ASHA told Aditi to eat vegetables, avoid spicy food, and take the IFA tablets; no other information was provided to Aditi about antenatal care, birth-preparedness, or her entitlements under the NRHM. During her pregnancy, Aditi went to the primary health center nearby for occasional check-ups but was not provided free antenatal care as specified under government norms. In the ninth month of pregnancy Aditi developed labor pain when she was at home. The ASHA came along with a private doctor and the delivery was conducted at home. The family did not receive the 500 rupees (US$10) for home deliveries as specified under the JSY. On the third day after her delivery Aditi complained of severe abdominal pain and fever. The family was not aware of women's entitlements to postnatal care or the requirement for ASHA visits in the postpartum period. No health worker had come to visit Aditi after her delivery. When the pain did not subside, Aditi's relatives arranged for a Marshall (jeep) to take her to the primary health center nearby. She died en route. Her baby died within a year of her death.  No health worker had come to inquire about or record Aditi's death.

The administrative division of work is such that health workers focus their delivery of services to women who are living in their matrimonial or husband's home. Therefore women who do not stay in their matrimonial or husband's homes often do not receive follow-up pregnancy health care. Women often moved between the matrimonial and maternal homes during pregnancy. Nearly 9 percent of all deliveries are said to occur in the maternal home.[225]

Information about pregnant women who moved between their matrimonial and maternal homes during pregnancy was lost within the system. ASHAs and anganwadi workers told Human Rights Watch that they did not normally attend to pregnant women residing in their mother's or relative's houses unless they stayed there permanently.[226] In at least two instances, health workers expressed their inability to record maternal deaths that had occurred within their coverage area because the women concerned died in their maternal homes.[227] 

Explaining the reasons why she did not report a maternal death that occurred in her village, ASHA Pooja P. said that Soumya S., the pregnant woman, had come to her aunt's house. When asked how that affected whether she should report the death, Pooja said, "I do not have to note down her name because I did not attend her case." She elaborated saying "Only bahus [daughters-in-law] of our village get registered. We are told in the training that we have to motivate only the bahus." She explained that the government did not provide money to ASHAs who followed-up on pregnant women for the period they lived in their mother's homes. But as she herself pointed out, "We get money if we motivate them for sterilization-150 rupees [US$3] for every case. It does not matter where the woman is [for sterilization]. I learnt all this from the training."[228]

A health official at the district-level told us that ASHAs do not get money for working with women from other villages. She said,

ASHA does not get money for bringing women from other villages. ASHA should take women for ANC [antenatal care], delivery, and PNC [postnatal care]. So if she cannot follow-up fully, then what is the fun in giving her money [incentive to ASHA]?[229]

Ratna R., an anganwadi worker who claimed that it was also her responsibility to record maternal deaths, stated that she had not recorded a maternal death that occurred in her village. According to her, in June 2008, Rohini R. delivered in the village health-subcenter, had convulsions soon after, and died. Ratna said, "This is Rohini's maikai's  [mother's house]village. So her death will not be noted here. We do not register women when they are in their maikai's." When asked whether someone else had registered her death she said, "In this case I do not think the ANM registered her death either. But I cannot be sure. But generally in such cases ANMs do not register the death."[230]

One stated rationale for this administrative policy is reduction of duplication of records and payments under the JSY.  However, to make recording of a maternal death dependent on whether the pregnant woman chooses to see through a pregnancy in a matrimonial home, maternal home, or elsewhere is an arbitrary distinction and linking access to maternal care based on whether a woman resides in her matrimonial home is discriminatory, not least on the basis of her marital status.  

A Hostile Reporting Environment

Health workers are reluctant to report deaths for fear, justifiably or not, of being singled out by government officials for punitive action.[231] Many activists and government officials feel that one of the key reasons for the non-implementation of the 2004 maternal death audit government order is a misunderstanding of the purpose of such an exercise-the fear that the audit seeks to find fault for maternal deaths rather than investigate systemic causes.[232] One district-level health official explained,

For those deaths reported through word of mouth, someone has to complain [for us to record and investigate it]. ASHAs must complain, but they are afraid it gives them a bad name. If we go to do an investigation, no one will come forward [to give us information].[233]

One district level official explained how overworked ANMs sometimes hide maternal deaths where they are not able to provide care as required of them. He said,

The tracking and monitoring [of maternal deaths] is very poor. How much can you expect one lady [referring to the ANM] to do? .... There is underreporting of [maternal] deaths. My personal experience has been that some ANMs hide deaths. They are busy-out for 10 days doing polio [administering vaccine]-they do not go to all of the villages. If there is a [maternal] casualty in this period, they do not report it.[234]

Referring to the fear of reporting maternal deaths within the public and private health sectors, Dr. Narendra Malhotra, the former president of the Federation of Obstetric and Gynecological Societies of India (FOGSI) said,

There is fear of reporting but we have to reassure them that the primary function of such reporting is to find the causes of death and save many more lives. The idea is to report one death and save other lives. That should be the key message.[235]

At the managerial level, the value of reporting and investigating maternal deaths is overshadowed by its perception as a performance indicator, causing officers to give low priority to improve such reporting systems. Sheela Rani Chunkath, formerly the health secretary of Tamil Nadu state, explained that "the attitude of the government generally is that high deaths means poor performance."[236]

Dr. P. Padmanaban, formerly a senior state health official in Tamil Nadu, and advisor to the Indian government on public health, described having raised maternal death reporting in a meeting organized for state health officials in Ahmedabad. He said, "No one was verykeen because it will expose many gaps in the health system. So no one wants to take it up. Performance linked issues-they feel if they report higher number of deaths, then they will be asked questions."[237]

Even though a 2004 government order instituting the maternal death audit system exists on paper, no government officials were able to give us detailed examples of inquiries and their outcomes.[238] One district medical officer cited an example of a maternal death in a primary healthcare center in September or October 2008. In that case an inquiry was held, resulting in the suspension of the concerned medical officer and nurses, but from the official version of the facts it appeared to be a clear-cut case of negligence. According to the official version, the ANM left the premises of the primary health center without informing the medical officer that a pregnant woman had just delivered a stillborn baby. As a result the mother was left unattended, developed post partum hemorrhage, and died within two hours.[239] 

Fear of inquiries, disciplinary action, and attacks by patients' relatives also create an environment that threatens free reporting. One staff nurse who had been suspended without an inquiry said, "I am naturally scared of reporting a death. I am only human. Over here if something goes wrong they will first suspend and only then will they find out if we even did anything wrong."[240]

Uncounted Deaths in Private Facilities

While some private facilities that are part of the Federation of Obstetric and Gynecological Societies of India develop registries that record maternal deaths,[241] district health officials do not collect information regarding maternal deaths from private facilities in Uttar Pradesh. Moreover, health officials from the district and state levels give conflicting reports about whether they have the power to collect such information.

A senior health official from the Directorate of Medical and Health Services told Human Rights Watch that district chief medical officers have the power to collect information on maternal deaths from private facilities. He said,

Power to register private clinics and hospitals is with the CMOs office. These clinics report to the CMO. CMO is the nodal officer and collects monthly information. [There is] also High Court ruling on private hospitals that need registration. So there is legal basis to collect information on maternal deaths.[242]

But district chief medical officers contradicted this, saying that they had no powers to collect such information.[243] When asked what information they had about deaths in private health facilities, one district chief medical officer said, "We do not come to know about private clinic deliveries ... Private hospitals do not have to report deaths to the CMO. What can we do? They just have to issue a death certificate."[244]

Acknowledging that "there is missing information [about] deliveries and deaths in private hospitals," an NRHM district program manager confirmed that they "do not have a system for reporting from [other] private hospitals." However, he clarified that the district health officials received some information from the two JSY-accredited private hospitals in every district. He recalled how this issue had been raised in a meeting with state-level officials at least twice recently. But even though state officials were apprised of the problem, they had not taken measures to improve the situation. Instead, state health officials felt that such decisions should be taken by district health officials. No such action was forthcoming from district health officials.[245]

Human Rights Watch visited four private facilities. Doctors from three of them stated that they were not reporting deaths because the district health officials did not ask them for this information.[246]

Uma U. 's death in a private health facility (Story as narrated to by Vignesh V. and Pratap P., husband and father-in-law of deceased mother) Uma U. , about age 20, was pregnant for the first time. In June 2008, in the ninth month of pregnancy, her husband, Vignesh V. took her to a government community health center for delivery. According to Vignesh they kept her there for about six or seven hours and when she did not deliver, the staff at the community health center recommended that she be taken to a private hospital. The staff identified the private hospital to which Uma should be taken and went with Uma's husband. In the private health facility, Uma had a surgery. Both mother and baby survived the surgery. She needed a blood transfusion and they arranged for it. But within five or six hours of the surgery, just as Uma began to receive the second bottle of blood, "her color changed and she started getting something like fits." When doctors were alerted they told Vignesh to take her to Kanpur to a bigger hospital. Alternatively, they asked for 25,000 rupees (US$520) for "ICU" facilities. Vignesh agreed to give them the 25,000 rupees. The doctors wheeled her in to another room and when she was brought out, she had passed away. According to Vignesh, all along the baby was kept in an incubator and the doctors had assured him that the "the baby was fine 95 percent." He claimed that as soon the mother died the doctors told Vignesh that the baby would not survive. Subsequently, the baby also died. Vignesh was told to take both the bodies and leave the hospital.

Several activists stated that poor monitoring and regulation of private facilities and doctors, especially those conducting unsafe abortions, results in unchecked deaths and complications from unsafe abortions. Further, they stated that unregulated unsafe abortion facilities make tracking abortion-related deaths more difficult; getting accurate information about abortion-related deaths is already difficult because these are often hushed up due to social stigma or misinformation that abortions are illegal even though India has a law allowing medical termination of pregnancies.[247]  As a result, they are often misclassified as suicides or murders, especially where the death is due to abortion by an unmarried pregnant woman.[248] Unsafe abortions contribute to roughly 10-12 percent of all maternal deaths in India.[249] Unless access to safe abortions is strengthened and abortion clinics are monitored as required under Indian law, the government cannot avert or track maternal deaths due to unsafe abortions.[250]

Discrimination

Caste-based discrimination not only adversely affects access to and utilization of health care but also affects reporting mechanisms. The effectiveness of the reporting mechanism is dependent on field-based health workers being in regular touch with families. If they are not in contact with families because of caste-based reasons, this not only affects the ability of the families to access health care but also adversely affects reporting.

Several surveys and studies have shown that women and girls from lower castes, especially Dalit communities, experience poor maternal health, underlining that the Indian central and state governments have yet to take adequate measures to provide equal access to health care for these communities.[251]

In most villages visited by Human Rights Watch, women were broadly aware of or had at least heard of the cash incentive for facility-based deliveries under the JSY. Many reported seeing health workers at least during immunization drives. In sharp contrast, women from a predominantly Kol (a group considered "untouchable") village in Chitrakoot district were completely unaware of the government health schemes.[252]Chunni C., one of the pregnant women residing in the village, told Human Rights Watch that she had not received any antenatal care, had not regularly seen an ANM or ASHA in her village, and seemed unaware of the JSY.[253] Laali L., another resident of the village, reported seeing the ANM for the first time 15 days prior to her interview with Human Right Watch, not before or after.[254]

Rathrani, an activist from Vanangana, a nongovernmental organization that has worked extensively in Chitrakoot district against caste-based discrimination, stated that there was one anganwadi worker, two ASHAs, and an ANM-all belonging to upper castes-in the vicinity but none of them visited the village because they considered Kols as "achooth [untouchable]."[255]Human Rights Watch spoke to the ASHA from the Biswakarma community, whose responsibility it was to visit the concerned village. She claimed that she was conducting regular visits.[256]

Rajdayya of the Dalit Mahila Samiti (Dalit Women's Society) explained how entrenched caste-based discrimination is in these areas. She said,

Biswakarma and Kol community follow untouchability. And Kol and Nayi community [follow untouchability] between them ... There are many areas where Biswakarma people will not go to the villages of Kol people.[257]

In their effort to dispel myths about caste hierarchies, the Dalit Mahila Samiti tried organizing a joint meal for women from different castes including Kols, Jamadars, and Biswakarmas. Rajdayya explained that she faced considerable resistance:

Kol community women did not want to come for the joint lunch [saying] that they knew their eyes would burst if they came for such a joint lunch. It was like paap [sin] for them. We had to force them to go to show them that it was not true.[258]

The Indian central government has declared the Kol community as a scheduled caste in Uttar Pradesh, acknowledging that this community is traditionally considered untouchable.[259]

In another village visited by Human Rights Watch, women belonging to the Chamar community described how the ANM from the Mishra community visited their village. Trishna T., a resident of the village belonging to the Chamar community said, "The ASHA and nurse [ANM] come only during polio [immunization] time. They come house-to-house then. Otherwise they do not come here. And they do not tell us anything-no meetings organized-nothing."[260]It was unclear whether the ASHA's or ANM's limited visits to the village were part of a larger systemic problem or motivated by caste-based discrimination. But when the women reported that even during the polio eradication drives, the ASHA or ANM came with someone from the Chamar community, they described what appears to be caste-based discrimination. Trishna said,

Even when they come they bring someone else who is a Chamar. He is the one who gives polio [drops]. The nurse is Mishra so she would not touch our children. They only come in the morning to write numbers on our houses and then will record in their registers whether polio [drops] was given.[261]

As is the case with the Kol community, the Indian central government has declared the Chamar community as a scheduled caste, acknowledging that they are also traditionally considered untouchable.[262]

The Tamil Nadu System of Investigating Maternal Deaths

Tamil Nadu has one of the lowest maternal mortality ratios in the country. Government officials and activists say that a number of socio-economic and political factors influence how health measures are implemented, and this in turn influences the maternal health situation in the state.[263] State health officials also have introduced several measures specifically aimed at reducing maternal mortality-"surveillance and audits of maternal deaths" being one of them.[264] In this section, we describe some of the ways in which the Tamil Nadu approach helps create constructive accountability. It documents some positive features of the Tamil Nadu system that can perhaps help inform similar processes elsewhere in the country. It does not advocate that the Tamil Nadu model is the best.[265]

The Tamil Nadu government has been documenting and analyzing maternal deaths for over a decade.[266] Health officials agree that their system has scope for improvement and is not foolproof.[267] Nevertheless, there are several positive lessons to be learned from Tamil Nadu's experiences in Theni and Dharmapuri districts.

First, all levels of the Tamil Nadu government-from the state to the village-recognize maternal mortality reduction as an important health priority and document and investigate maternal deaths as an intervention towards this end. Sheela Rani Chunkath, the former health secretary who is considered one of the pioneers of the Tamil Nadu initiative, said,

Reducing maternal mortality is like a complicated management problem...you need to know how to manage scarce resources-limited number of anesthetists, surgeons, blood banks. For this you have to identify gaps. I have found that conducting maternal death reviews not only sensitizes health staff but also helps in this management process.[268]

Second, investigations in the Tamil Nadu system have a clear purpose: identifying health system gaps that can be improved upon. Almost all state and district health officials that Human Rights Watch spoke to who had participated and overseen the Tamil Nadu system of investigating deaths stated that the initiative was started with clear objectives-to change the idea that maternal deaths are "normal" or "fate," redirect responsibility to the health system instead of blaming families or health workers for maternal deaths, and develop "actionable points to change the health system."[269] 

Third, the maternal death reporting system in Tamil Nadu covers all pregnancy-related deaths irrespective of the medical cause of death and place of occurrence.[270] Since the medical cause of death is difficult to ascertain in some cases, relatives and health workers are encouraged to report any pregnancy-related death. Alphonse Mary, the Maternal and Child Health Officer of Dharmapuri district, who conducts inquiries into maternal deaths said, "I get all kinds of cases. Last month I got a case of a pregnant woman who drank poison and died."[271]

The investigation is undertaken at the district level involving officers who have powers to bring about programmatic changes within the health system. District level officials with judicial powers (district collectors) exercise their powers to ensure that private facilities also participate in district-level maternal mortality review meetings. Alphonse Mary said,

Our collector [district level official with judicial powers] has issued summonses to doctors from a private hospital to attend the maternal mortality review meeting because they were not coming. They [doctors from the private facility] have to come .... They had admitted a woman needing a blood transfusion when they knew they did not have the facility. [They] should come and explain why.[272]

Fourth, the Tamil Nadu government has tackled head-on the problem of gross underreporting of maternal deaths. They struggled with underreporting for several years when the system was initially started in the 1990s but there has been incremental improvement in their reporting system annually.[273] Officials stated that a series of awareness programs and training for health workers coupled with a system of multiple reporting of deaths improved their reporting systems.[274] Dr. Kolanda Swamy, who has overseen reporting of maternal deaths for many years in his capacity as deputy director for health services, said,

Encourage multiple reports [of a maternal death]. Anyone should be allowed to report. Not just health workers. For health workers it is mandatory. Appreciate people who are reporting maternal deaths. Censure those who are not reporting. It is much worse for our health workers to not have reported a death at all than to have reported it and come for a maternal death review meeting.[275]

Dr. Iyyannar, deputy director of health services who oversees health concerns in the Dharmapuri district of Tamil Nadu, emphasized the importance of involving relatives to ascertain the actual cause of death. He said,

Sometimes doctors will give positive findings even within the best system. They will rewrite the case sheets. But relatives will tell you exactly what happened and you can get to the bottom of what happened. Simple things like noting time of entry to the hospital and time of referral [can be rewritten]. But attendants coming with the patient will know. And they will give you the exact story.[276]

When asked whether the Tamil Nadu government initiated punitive action against health workers for maternal deaths, government officials stated that in a majority of cases they found that systemic gaps and faults were responsible for the maternal death. "Before blaming doctors and nurses, we have to ensure that the entire system that supports them is working," said Chunkath. Explaining how there was a fine line between negligence, unethical practice, and systemic failure, Dr. Swamy described how they tried to balance these concerns to determine whether to initiate action against individual officers. "After having all the facilities, training and knowing how to handle the case, if a doctor evades responsibility-for example-by refusing to come on time or purposely referring the case to another hospital, then we can say clearly it is unethical practice," said Dr. Swamy.[277]  Similarly, Dr. Iyyanar and Alphonse Mary shared examples of cases where they had initiated action against doctors or nurses who they felt had clearly evaded responsibility.[278]

Failure to Use Appropriate Indicators

Institutional Deliveries as an Inadequate Indicator of Progress

It is not enough to design and implement apparently worthwhile activities. We have to make sure they work on the ground. Process and output indicators are especially useful in this regard because they provideinformation not only for the final evaluation, but also for ongoing management and improvement of program components.

- Deborah Maine et al., June 1997.

Constant monitoring of the progress of maternal health interventions based on appropriate indicators is critical to evaluating the effectiveness of programs aimed at reducing maternal mortality. After the introduction of the JSY in 2005, both the Indian central and Uttar Pradesh state governments have used increased institutional deliveries (births in healthcare facilities) as a key measure of progress.[286] The Indian central government states that from mid-2005 to March 2009 nearly 20 million women had benefited under the JSY across India.[287] However, using institutional deliveries as a proxy indicator of progress is problematic for several reasons.

The Uttar Pradesh state government monitors the success of the JSY by setting and achieving "targets" for the number of facility-based deliveries, without monitoring or ensuring that such deliveries are in fact safe and include women who develop pregnancy-related complications.[288] The Uttar Pradesh government fixes annual targets for institutional deliveries both at the state and district levels.[289] Several district chief medical officers confirmed that such targets are set while preparing district NRHM plans for the following financial year, and are based on population and birth rate.[290] "There is a target of about 24,000 institutional deliveries under JSY. We are 10 percent over target," said the Unnao district chief medical officer. The Chitrakoot district additional chief medical officer reported that they had achieved 92 percent of their target by March 2009. Saying that these are increased "exponentially each year," the chief medical officer of Allahabad elaborated that, "Last year it was 13,059. This year it is 35,000. Next year it will be 61,000."[291]

The Uttar Pradesh government monitors achievement of targets not through the actual number of safe deliveries, but through the number of beneficiaries.[292] Whenever a JSY cash incentive is paid to a woman who delivers in a health facility, she is counted as a "beneficiary," and her delivery is counted towards meeting the target.[293]

The accuracy of the reported number of facility-based deliveries in Uttar Pradesh is questionable. The pressure to demonstrate increased institutional deliveries has resulted in spurious payment practices in many cases, skewing the JSY data. Several rural women reported that ASHAs or ANMs had approached them to show their deliveries as facility-based even though they were home-based.[294]

Even assuming that the JSY data on the number of facility-based deliveries is accurate, this data alone does not throw light on the extent of impact of the JSY without information and analysis of adverse outcomes such as maternal deaths and severe morbidities. In what appeared to be an extreme case, in March 2009 health workers had paid out the JSY cash incentive to a the family of R., a 26 year-old woman from Azamgarh district, counting her as a beneficiary under the JSY scheme. R. had developed complications during delivery in a health facility and died.[295] "Even though she died, the fact that she was counted as a beneficiary would be reflected in government's records as a successful institutional delivery," said Jashodhara Dasgupta, an expert on maternal health and women's rights from Uttar Pradesh.[296]

While the government can show that a woman delivered in a health facility, they are not tracking whether every registered pregnant woman actually delivered safely, developed complications, or died during the post-partum period.[297] For example, if there are 50 facility-based deliveries and 10 women eventually died due to post-partum complications, the latter fact may never show up in government records.

Neither of the state agencies responsible for family welfare and health collects data on maternal deaths to analyze JSY's impact. A senior official described the JSY as a strategy for reducing maternal mortality, but in the same breath asserted that there was no connection between its monitoring parameters and maternal deaths. She said,

We [the Directorate of Family Welfare] are not collecting any information on maternal deaths. But we want this information ... it is in the pipeline. JSY and maternal deaths are two separate things. Under JSY we only do monitoring of institutional delivery. 'Maternal deaths' is separate and falls under [the] health directorate – this information comes to the CMO [chief medical officer, district-level official].[298]

Contrary to what the official from the family welfare directorate claimed, officials from the Directorate of Medical Services and Health (health directorate) stated that they were not concerned with maternal health-related data including deaths.[299]  One officer said, "Maternal health, maternal deaths, anything connected to the mother-all this comes under the family welfare directorate. It does not come under us."[300]

There is also no clear breakdown of the number of institutional deliveries by type of care-basic care or comprehensive emergency care. When sharing the data collated under the JSY, district medical officers showed figures for facility-based deliveries. In some cases, these were broken down by place of delivery-home or health facility (sub-health center, primary health center, community health center, or district women's hospital).[301] But they had no concrete information about the percentage of such deliveries that successfully addressed pregnancy complications.[302] Many health staff in community health centers reported that they largely conducted "normal deliveries."[303]

Former senior state health officials, doctors, and activists say that it is precisely for this reason that the extent of JSY's impact on maternal mortality is unknown.[304] A former senior health official in Uttar Pradesh was doubtful whether the women who actually needed emergency medical care for pregnancy-related complications were in fact benefiting from the JSY scheme.[305] Government NRHM review reports raise similar concerns that remain unaddressed.[306]

Dr. Chandravati, former professor of gynecology at the medical college hospital in Lucknow and an advisor to the Uttar Pradesh state health department, said that the JSY had not yet resulted in "an identifiable decrease in cases of complications and deaths." She felt that it would possibly take more time to show results, and cautioned that gaps in the scheme would need to be addressed to achieve progress. She expressed concern about the scheme, saying that "lots of gaps are there-the facilities are not upgraded and suddenly the load on these institutions has increased."[307]

The UN Special Rapporteur on health expressed similar concerns after his visit to India in 2007. He noted that the Indian authorities have successfully managed to increase the number of women delivering in health facilities, "[b]ut, in many cases, the range and quality of services offered in those facilities has been seriously neglected. In short, the supply-side has received too little attention."[308] He went further to make the important distinction between provision of facility-based care and access to life-saving treatment: 

The focus has been on increasing institutional delivery -but institutional delivery does not always provide access to life-saving care, such as emergency obstetric care, and therefore cannot be regarded as a proxy for access to life saving care."[309]

Activists repeatedly pointed out to Human Rights Watch that the JSY's underlying assumption, that institutional deliveries are safe deliveries, is problematic.[310] Institutional deliveries cannot by default be treated as a measure of deliveries by a skilled birth attendant.[311] The presence of a skilled birth attendant at delivery is associated with better delivery outcomes, including reduction in maternal deaths.[312] But such association is considered plausible only where a trained attendant authorized to perform life-saving functions is supported by a performing health system that can provide life-saving interventions in a timely manner.[313] According to a joint statement issued by the World Health Organization (WHO), the International Confederation of Midwives (ICM), and the International Federation of Gynecology and Obstetrics (FIGO), the definition of skilled care itself requires that an "accredited and competent" healthcare provider has at her disposal the "necessary equipment and the support of a functioning health system, including transport and referral facilities for emergency obstetric care."[314] A skilled birth attendant is a healthcare provider who is "trained to proficiency" not only in the skills needed to manage "uncomplicated" cases, but also to identify, manage, and refer complications.[315] 

Concerns about the quality of maternal health care, availability of health workers with midwifery skills, and the level of support afforded to such health workers to perform life-saving interventions indicate that the government's assumption that all institutional deliveries are safe is not well supported.[316] Several government officials and doctors in Uttar Pradesh consistently maintained that they did not have the facilities to meet the "demands" for institutional delivery. One senior health official said, "JSY has opened up the gates for institutional deliveries ... Quality is lacking. Once you create the demand, then your facilities have to be ready."[317]

Poor Monitoring of Emergency Obstetric Care Indicators

Measuring maternal mortality ratios is one way of assessing trends and progress in maternal mortality reduction. However, given that such measurement is contingent upon periodic surveys and a strong vital registration system,[318] experts have developed supplementary approaches for measuring progress. The UN process indicators and guidelines, a set of six indicators, are based on the understanding that certain types of obstetric services must be made available and used to reduce maternal mortality.[319] The six indicators address the minimum required number of basic and comprehensive emergency obstetric care facilities for a given population, their geographical distribution, minimum proportion of births that should occur in basic and comprehensive emergency obstetric facilities, whether women with pregnancy complications were in fact treated in emergency obstetric care facilities, acceptable proportion of births through cesarean sections, and the number of deaths among women with pregnancy complications admitted to facilities equipped with emergency obstetric care.

Even though these indicators are neither binding nor adopted by the UN General Assembly, they were initially developed by UNICEF and have thereafter been used by other UN agencies like the UNFPA and the WHO and are commonly referred to as the UN process indicators. Information based on these indicators reveals whether health facilities for basic and comprehensive emergency obstetric care exist, are distributed in a useful fashion, are used by women, and are used by women who develop obstetric complications. They have been issued with detailed guidelines and minimum norms for each indicator.[320]

The Uttar Pradesh government's routine monitoring of maternal healthcare programs does not take into account the UN process indicators.[321]  

Even though periodic surveys like the National Family and Health Survey (NFHS) and District Level Household and Facility Survey (DLHS) provide useful information about the status of maternal health in the country, these do not provide information on key indicators such as whether the need for emergency obstetric care was met in all cases and the proportion of maternal deaths among women with obstetric complications admitted to facilities.[322]

The new NRHM Health Management Information System (HMIS) seeks to fill this gap to some extent by including data on obstetric first aid and access to first referral units or health facilities equipped with emergency obstetric care. But the HMIS still does not include information on all the recommended UN process indicators.[323] Further, government advisors from the National Health Systems Resource Center overseeing implementation of the HMIS in different states say that they have experienced considerable difficulty in rolling out the system in Uttar Pradesh, referring to it as a "problem state."[324]

The Indian central government is also conducting a new Annual Health Survey in several states including Uttar Pradesh, tentatively from 2009.[325] The Registrar General's office-the office coordinating the study- did not provide additional information to Human Rights Watch when asked in May 2009, stating that the survey was yet to be finalized.[326] It remains to be seen whether the survey will generate reliable information on maternal health or be in accordance with recommended UN indicators and guidelines.

Poor Long-term Monitoring

Civil Registration and Maternal Mortality

Civil registration is essential for continuous and long-term monitoring of the progressive realization of women's right to health.[327] A strong civil registration system recording vital events such as births and deaths, including the cause of death, has immense implications for good public health policy and decision-making.[328] Almost all developed countries have a reliable national civil registry recording vital events.[329] Reliable civil registration data is almost never available in low or middle income countries.[330]

In the context of maternal health, the Indian government itself has acknowledged that the "absence of reliable estimates" of maternal mortality makes the process of reducing it "both difficult and complex."[331]  More recently, the UN Special Rapporteur on health, after his 2007 visit to India, said, "[t]here is no effective, reliable and comprehensive civil registration system for accurately reporting births and deaths in India. There is evidence that women are silently dying in childbirth and during pregnancy."[332]

The UN Millennium Project Task Force on Child and Maternal Health has called for strengthening civil registration as a "critical investment for reaching the [Millennium Development] Goals." The Task Force stated:

While modeling and population-based surveys can augment our understanding of general levels and trends, they are not a substitute for strong, country-owned vital statistics and civil registration systems ... This task force seconds the call for information, starting with a simple accounting of who is born and who dies, as a critical crosscutting investment necessary for reaching the Goals.[333]

In addition to assisting in the long-term monitoring of maternal mortality, civil registration also helps monitor the progressive realization of many other rights. A robust civil registration system with universal birth registration can help monitor early and enforced marriages. Early marriage and child-bearing have a direct impact on the sexual and reproductive health of girls and women.[334] The median age of marriage in India is 17.2 years, below the legal minimum of age 18.[335] 2007-2008 data from rural Uttar Pradesh shows that early child-bearing continues to remain a problem.[336]

Joint Failure of the Indian Central and Uttar Pradesh Governments

Recording vital events such as births and deaths is mandatory under Indian law under the Registration of Births and Deaths Act, 1969.[337] The duty to implement the law rests both with the Indian central and the concerned state government.[338] The 2000 Indian National Population Policy aims to achieve 100 percent registration of all vital events-births, deaths, and marriages by 2010.[339]

Even though registration of births and deaths is mandatory, civil registration in India is poor. The Central Bureau of Health Intelligence reports that an estimated 26 million births and 9 million deaths occur in India every year, of which only 53 percent of births and 48 percent of deaths are registered. Further, about 10 million births, constituting roughly 25 percent of the world's unregistered births and about 4 million deaths go unregistered annually.[340]

Uttar Pradesh has the worst civil registration record in the country. It is the only state to be repeatedly singled out for poor performance in the annual National Conference of Chief Registrars of Births and Deaths.[341] The latest report of the Indian government on civil registration released in March 2009, covering a backlog of nearly ten years between 1996 and 2005, reveals that the Uttar Pradesh state government has not submitted regular and reliable information on births and deaths to the Indian central government since 1996. Despite the prolonged delay in publishing the report, the central government still has no reliable data on civil registration from Uttar Pradesh.[342]

The reasons for such poor implementation of the civil registration system have been well documented and studied through government reports, notably reports issued pursuant to  the Birth and Death Registration Act and the Central Bureau of Health Intelligence 2007 report on mortality statistics in India.[343] The Indian central government has noted that "[registration] functionaries at all levels do the work of registration in addition to their other normal duties in an honorary capacity generally, without any incentive," and has concluded that "[this] is why the work of registration, preparation and submission of statistical returns do not get due attention and priority." Such lack of attention and priority manifests itself in many ways including poor budget allocation, and poor training and supervision of staff involved in registering births and deaths.[344]

Officials at Registrar General's office agree that poor priority for such data is the primary problem and gave some examples of how this is reflected in practice. "A lot of lower level officials [for registration] are not appointed," said one official. As of 2006, a mere 9,000 panchayat (local village council) secretaries were overseeing civil registration in 52,000 panchayats in Uttar Pradesh.[345] Local activists say the vacancies have been slowly filled since 2008.[346] Even where registering authorities have been appointed, however, they often have not been able to carry out their duties because they do not have the required forms and stationery for the registration.[347] Further, coordination meetings between block level officials, district level officials, and the state level inter-departmental coordination committee for civil registration are too seldom conducted. "Even if they meet and take even some small decisions-those decisions are not implemented," said an official from the Registrar General's office.[348] 

Despite repeated directives issued to the Uttar Pradesh government at the annual conference, and the host of commitments made by the Uttar Pradesh Chief Registrar to improve the situation, many districts of Uttar Pradesh continue to have "zero" birth and death reporting and registration.[349] One official from the Registrar General's Office stated that the Uttar Pradesh government had almost consistently defaulted on providing timely annual proposals for implementing the civil registration system.[350] Even when the Indian central government has released funds for implementing proposals, "[i]t is a perennial problem with UP that they say they cannot give us a consolidated statement of expenses [showing utilization of funds for civil registration]-and then finance here cannot release funds on time. This has been happening almost consistently since 2001," said the official.[351] 

Lamenting how things had not changed, one activist described his experience of working with Uttar Pradesh state government officials. "They [Directorate of Medical and Health Services] had their birth and death registration manuals in a big hall. All of them were piled up instead of being distributed to the Gram Vikas Adhikaris [village development officers]," he said. Seeing that the manuals were not being used, they approached the director general seeking permission to distribute the manuals. But this turned out to be an impossible task. "[T]hey raised [a] hundred objections. 'These are issued by Government of India -this and that-we cannot hand it over to you,' said the government" he explained. Eventually "they just went to waste."[352]

In February-March 2009 Human Rights Watch found that in many cases that registrars, including officials and healthcare workers, who are suppose to notify the authorities about births and deaths knew little about their duties under the law. None of the families, staff nurses, ANMs, or ASHAs were even aware that they were supposed to report maternal deaths to the registrar in their village, let alone use a particular format.[353] Activists and families stated that not all gram sabhas (cluster of villages governed by a village council) had a burial or cremation ground maintained by the village council for different religious communities.[354] Where the cost of transporting the dead body to a burial or cremation ground is high, and the family cannot afford it, they dispose of the body in their fields.[355]

Activists say that those who are listed as registrars on paper often exhibit little or no awareness of their appointment as registrars, much less an understanding of their obligations under the law. ANMs and superintendents of community health centers whom Human Rights Watch interviewed were unaware of their appointment as registrars and did not have the prescribed forms and registers to implement the Registration of Births and Deaths Act.[356]

In practice, panchayat mitras (literally, friends of the village council) assist village development officers in discharging their duties as registrars. One such panchayat mitra said,

 The gram vikas adhikari finds out about a death if someone from the family comes and makes an application .... If the family does not come then obviously we cannot know about the death. And most families do not come. Out of 100 deaths, about 20 people will come and make an application – only those with bheema [insurance] or other money or property will come and make an application.[357]

Another panchayat mitra said, "We find out about deaths if the pradhan (village head) goes to the villager's house to offer condolence [after the death]. Then this information goes to the gram vikas adhikari."[358]

None of the families with whom Human Rights Watch spoke who had suffered maternal deaths knew they needed to approach the ANM, the village development officer, or the panchayat mitra  to register the deaths.[359]

 The Uttar Pradesh government has issued an executive order (government notification) making the Medical Cause of Death Certification scheme applicable in many hospitals.[360]  However, the scheme is being implemented poorly. One official from the Registrar General's office stated that the "[s]ituation in UP [Uttar Pradesh] is very grim," pointing out that only 0.7 percent of the total registered deaths were medically certified. Until 2004 the Uttar Pradesh government had submitted data under the MCCD scheme only for four hospitals from the entire state even though their government notification covered more hospitals. "From 2005 there is no data at all," he concluded.[361]

As mentioned before, civil registration is the shared responsibility of both the Indian central and the Uttar Pradesh state government under the Indian Constitution.[362]Despite this shared responsibility, national authorities are doing very little to intervene and set right the problem.