Implementation research on kangaroo mother care, Bangladesh

Abstract Objective To describe the implementation, coverage and performance of the national kangaroo mother care programme in Bangladesh. Methods Kangaroo mother care services for clinically stable babies with birth weight under 2000 g were set up in government-run health-care facilities in rural and urban areas of Bangladesh. Each facility provided counselling on kangaroo mother care, ensured adequate nutrition, and followed up mothers and babies. We studied implementation of the programme from January 2016 to March 2020 using data from the national database. We tracked the number of eligible babies enrolled and their outcomes, mortality and post-discharge follow-up. Findings The numbers of kangaroo mother care facilities increased from 16 in 2016 to 108 in 2020. Over the 4-year period 64 426 babies weighing under 2000 g were born in these facilities,  6410 of whom received kangaroo mother care. The quarterly percentage of eligible babies receiving kangaroo mother care increased from 4.7% (37/792) during the first quarter to 21.7% (917/4226) during the last five quarters of the programme. Deaths of babies receiving kangaroo mother care showed a downward trend over the study period. The overall mortality was 1.2% (77/6410), with large quarterly fluctuations in mortality. Post-discharge follow-up was low and only 15–20% of babies received four follow-up visits. Conclusion Implementation of kangaroo mother care interventions is feasible in low-resource settings. Such care has the potential to reduce mortality among low-birth-weight and premature babies. Challenges include low coverage, expanding the programme to the community and strengthening the monitoring system.


Introduction
Prematurity and its complications are one of the leading causes of neonatal deaths worldwide. Every year 15 million babies in the world are born prematurely. 1 Low birth weight, defined as babies born with birth weight below 2500 g, frequently coexists with prematurity. Lower birth weight is associated with higher mortality, especially below birth weight 2000 g. 2 Kangaroo mother care is a low-cost, low-technology intervention, defined as "a standard, protocol-based care system for preterm and/or low birth-weight newborns based on skin-to-skin contact between the newborn and the mother or the caregiver. " 3 More than 200 hospital-based studies have compared incubator care with kangaroo mother care and determined that kangaroo mother care is effective in maintaining adequate temperature, reducing hospital-acquired infections, encouraging exclusive breastfeeding and promoting neonatal weight gain. 3 Estimates show that kangaroo mother care could prevent up to 450 000 preterm deaths each year if near universal coverage of the practice is achieved. 4 Although success of kangaroo mother care has been widely reported from hospitalbased interventions, 5 there are fewer published studies of the implementation of kangaroo mother care at a national level.
Bangladesh, a lower-middle-income country, has made significant progress in reducing child mortality. However, neonatal mortality is still high and the burden of prematurity is increasing. Bangladesh is one of the top 10 countries with the largest number of preterm births. 6 An estimated 31% of all neonatal deaths in Bangladesh are due to prematurity and its complications. 7 The World Health Organization (WHO) recommends kangaroo mother care for routine care of newborns weighing below 2000 g at birth to be initiated in health-care facilities as soon as the newborn is clinically stable.
In 2013, the Government of Bangladesh declared its commitment to introduce and scale up kangaroo mother care in health-care facilities, with continuation of care at home. To end preventable child deaths by 2035, the government renewed its declaration in Child survival call to action: a promise renewed. 8,9 One of the aims was to reduce the under-five mortality rate to 20 per 1000 live births by 2035 through implementation of different strategies including kangaroo mother care for low-birth-weight babies. 8 A technical subgroup was formed under the guidance of Bangladesh's national technical working committee for newborn health. In 2014, the committee developed the national guideline on kangaroo mother care in line with the WHO kangaroo mother care guideline. 10 The national guideline was designed for health-care providers and managers and focused on the introduction, expansion and strengthening of kangaroo mother care practices. In 2015, kangaroo mother care was included as one of the key strategies in the Bangladesh Every Newborn Action Plan. 11 The main activities planned were: (i) ensuring counselling Kangaroo mother care, Bangladesh ANM Ehtesham Kabir et al.
on kangaroo mother care in healthcare facilities; (ii) providing follow-up kangaroo mother care services in the community; (iii) establishing centres of excellence for kangaroo mother care in secondary and tertiary level facilities; and (iv) setting targets for initiation of kangaroo mother care. In 2016, kangaroo mother care was incorporated as a priority newborn health intervention in the implementation plan of the 4th Health, Population and Nutrition Sector Programme of the Ministry of Health and Family Welfare. 12 We describe the first 4 years of implementation of the kangaroo mother care programme in Bangladesh. We aimed to demonstrate that implementation of the programme in a low-resource country is feasible and can reduce neonatal mortality. We also compiled lessons learnt to inform further scale-up of the programme in Bangladesh and other low-and middle-income countries with a similar socioeconomic context.

Setting
The government introduced kangaroo mother care nationwide as a part of its comprehensive strategy to reduce neonatal mortality. 12, 13 The target was to establish kangaroo mother care at 20% of public health facilities by 2016 and at 50% of public health facilities by 2020. 9 Kangaroo mother care services were planned for subdistrict health complexes and higher-level facilities (district hospitals and medical college hospitals) in both urban and rural areas. Under the programme, a dedicated space (corner) for kangaroo mother care was established in each designated health facility. Although 142 kangaroo mother care corners were originally planned, by 2020 only 108 corners remained functional under the supervision of the Directorate General of Health Services. In addition, 22 kangaroo mother care corners were established under the Directorate General of Family Planning, but we only evaluated kangaroo mother care facilities under the Directorate General of Health Services.

Implementation
In 2014-2015, as a part of the implementation plan for kangaroo mother care, the government developed a national trainer pool for health-care professionals. Train-ing programmes included a 3-day training on kangaroo mother care focusing on practical demonstration, learning by doing and facility visits for direct observation of kangaroo mother care services. The programme also included staff training on recording and reporting of data. The national newborn health programme implemented a social and behavioural change communication campaign to create awareness in the community via printed materials (such as leaflets, posters, flipcharts and booklets), television, radio and social media platforms.
In 2016, kangaroo mother care pilot programmes were initially introduced at 21 health-care facilities, with 16 remaining functional. Health-care providers were trained by a pool of physician trainers to provide kangaroo mother care services; record-keeping and reporting systems were established; and managers and community health workers were prepared. Each service centre provided counselling to mothers on kangaroo mother care, ensured adequate nutrition through counselling about breastfeeding, provided support services to other centres and followed up mothers and babies (Fig. 1). The programme included four planned follow-up visits after discharge at weekly intervals. All eligible babies were managed according to standard facility protocols along with kangaroo mother care services. The details of the training and implementation have been described previously. 9 The online kangaroo mother care data set was also launched in 2016. Data are entered by trained personnel within the facilities and periodically monitored by the Directorate General of Health Services for completeness and accuracy. Staff from the national newborn health programme evaluated kangaroo mother care data every month and acted on underreporting facilities. Kangaroo mother care services were monitored quarterly by government personnel through the national trainer pool, whereby professionals and programme personnel visited facilities and conducted onsite monitoring. This reference group provided expert opinions to service providers facing difficulties in implementing kangaroo mother care. Progressively, kangaroo mother care was introduced formally into the curricula of medical and nursing professionals. In parallel, various professional organizations of paediatricians, obstetricians, neonatologists and nursing colleagues endorsed the universal practice of kangaroo mother care. 9

Fig. 1. Pathway of care for babies born in designated kangaroo mother care centres, Bangladesh
Low-birth-weight babies delivered in or admitted to designated health-care facilities Babies are assessed for eligibility for kangaroo mother care. Eligible babies are clinically stable newborns with birth weight < 2000 g Initiation of kangaroo mother care in the health-care facilities. Care includes keeping mother and baby together (rooming-in) and moving them to dedicated kangaroo mother care space (corner) of facility Facility sta provide advice on kangaroo mother care position and nutrition (breastfeeding or alternative feeding method). Sta are also responsible for monitoring of care, record-keeping and preparing mothers for discharge Facility sta counsel mothers at discharge and create schedule for follow-up monitoring In the facility: regular follow-up of mothers and babies every week until 40 weeks of corrected gestation or resolution of any danger sign Complete data submission to central database In the community: four visits at weekly intervals scheduled for follow-up of mothers and babies during home visits by community health workers

Research
Kangaroo mother care, Bangladesh ANM Ehtesham Kabir et al.

Evaluation
We used quantitative methods to evaluate the implementation of the kangaroo mother care programme from January 2016 to March 2020. Our primary data source was the national data set for kangaroo mother care within the DHIS2 health management data platform (Health Information Systems Programme, Oslo, Norway). The data set includes both operational and outcome variables such as number of facilities having dedicated kangaroo mother care services, number of trained personnel, number of live births, number of babies with birth weight below 2000 g receiving the care, outcomes of babies who received the care, number of deaths, and follow-up rates of mother-baby pairs receiving the care. We used four outcome measures: (i) number of eligible babies enrolled in kangaroo mother care services; (ii) outcomes of babies receiving kangaroo mother care; (iii) mortality among babies receiving kangaroo mother care; and (iv) post-discharge follow-up rates. We included all clinically stable liveborn babies with birth weight below 2000 g at designated kangaroo mother care facilities. A clinically stable baby was a baby with normal respiratory and heart rate, normal oxygen saturation and without major congenital malformations. We report descriptive statistics with frequencies and percentages plotted using Excel version 2013 (Microsoft Corp. Redmond, United States of America). We also monitored the percentage of babies receiving kangaroo mother care who were: discharged as per the protocol, discharged on request, discontinued kangaroo mother care or died. We calculated mortality as the percentage of babies receiving kangaroo mother care who died within the facility while receiving kangaroo mother care. We calculated follow-up rates as the percentage of babies receiving each of the four planned weekly follow-up visits after discharge.
Approval was taken from the Directorate General of Health Services for analysis of data from the database. No individual health records were accessed or reviewed.

Results
The number of functioning kangaroo mother care corners was increasing over the study period, from 16 in 2016 to 108 in 2020. Among them 29 (26.9%) corners were in urban areas and the remaining 79 (73.1%) were in rural areas. The location of facilities was as follows: 79 (73.1%) in subdistrict health complexes, 21 (19.4%) in district hospitals and eight (7.4%) in medical college hospitals.
Multiple development partners provided technical assistance to the government personnel at those sites. A total of 1196 health-care providers were trained from 2018-2020 in 60 batches, among whom 828 (69.2%) were nurses, 228 (19.1%) were midwives, and 140 (11.7%) were doctors.
A total of 3 615 453 babies were born in the kangaroo mother care facilities during the study period, of whom 64 426 (1.8%) had birth weight less than 2000 g (Fig. 2). Table 1 shows the quarterly data set for the study duration. Fig. 3 shows the enrolment of babies to kangaroo mother care. Over the period January 2016 to March 2020, 6410 babies with birth weight under 2000 g received kangaroo mother care in the facilities. In the first quarter of the implementation 4.7% (37/792) of eligible babies received kangaroo mother care. However, the percentage decreased to 1.4% (53/3792) in April to June 2017, despite the higher number of eligible babies identified. We can attribute the decrease to multiple factors such as poor compliance, lack of familiarity with protocols and missing cases. From April 2018, after implementation of corrective measures, such as more stringent continuous monitoring and training of additional providers, the situation improved and in the last quarter of the study 21.7% (917/4226) of eligible babies received kangaroo mother care. Fig. 4 shows the trend in outcomes of babies receiving kangaroo mother care. The proportion of eligible babies receiving the care and discharged as per the protocol showed an upward trend, reaching 50-60% in the last three quarters. The remaining 40-50% of Source: We extracted data from the government's DHIS2 health management data platform.

Research
Kangaroo mother care, Bangladesh ANM Ehtesham Kabir et al.
babies were discharged at a later period after resolution of problems such as difficulty establishing breastfeeding or supplemental feeding, hypothermia and electrolyte disturbances. The proportion of babies who received kangaroo mother care but later discontinued care or were referred to higher-level facilities such as special care newborn units or neonatal intensive care units continued to drop from 30% during April-September 2017 to below 10% during the last eight quarters of the study period. The overall mortality among the babies receiving kangaroo mother care was 1.2% (77/6410), with a downward trend over the study period (Fig. 5). The data showed large quarterly fluctuations in mortality, sometimes due to lower numbers of deaths and sometimes due to more babies receiving kangaroo mother care during the first few quarters of the programme. Fewer fluctuations were found during the later part of the study as the number of enrolments increased. Fig. 6 shows the follow-up rates for babies who received kangaroo mother care. Only the first of the four follow- Note: We included all clinically stable liveborn babies with birth weight below 2000 g at designated kangaroo mother care facilities. A clinically stable baby was a baby with normal respiratory and heart rate, normal oxygen saturation and without major congenital malformations. Source: We extracted quarterly data from the government's DHIS2 health management data platform. up visits was sustained with 30-50% of babies receiving one follow-up visit until the end of the evaluation period. The percentage receiving second, third and fourth follow-up visits was suboptimal, although it improved from a baseline of 0% to around 15-20% in the last quarters of the study period.

Discussion
We report the implementation and initial evaluation of the kangaroo mother care programme in Bangladesh. We noticed a decreasing trend in mortality among babies with birth weight below 2000 g receiving kangaroo mother care in the designated facilities, with an overall mortality of 1.2% over the study period. This value is significantly lower than the historical neonatal mortality (deaths before or at 28 days of life) of 19% among babies born weighing less than 2500 g in Bangladesh. 14 Of note, the kangaroo mother care programme Year, month % of babies receiving kangaroo mother care  enrolled babies with birth weight less than 2000 g -a more vulnerable group that is expected to have higher mortality than babies weighing between 2000 g and 2500 g. However, we counted deaths that occurred while in the kangaroo mother care facilities and not neonatal deaths after discharge from the facilities. Other parameters showed a modest improvement. For example, enrolment in kangaroo mother care, although improved, reached 20%. Follow-up after discharge -a key element of the WHO kangaroo mother care guidelines -was low, especially for the second, third and fourth follow-up visits.
In an earlier randomized controlled trial of community-based kangaroo mother care among babies of all birth weights in Bangladesh, no overall difference in mortality was found although there was a reduction in mortality among babies with birth weight less than or equal to 2000 g. 15 Among newborns randomized to community kangaroo mother care, 77.4% (1565/2022) of infants received any skin-to-skin care, with only 23.8% of 1985 infants receiving skin-to-skin care for more than 7 hours per day in the first 2 days of life. Similarly, a large randomized controlled trial studying community initiated kangaroo mother care in rural India reported a 30% increase in neonatal survival (hazard ratio: 0.7; 95% confidence interval, CI: 0.51-0.96) and a 25% increase in infant survival at 6 months (hazard ratio: 0.75; 95% CI: 0.60-0.93). 16 We add to the literature on kangaroo mother care and demonstrate that kangaroo mother care, implemented in low-resource settings at national level, can also result in a progressive reduction of mortality among neonates.
Our study has some limitations. The data were incomplete, especially during the initial days of the study, due to logistical challenges of collecting and monitoring the data. Even in the best performing quarter, about four fifths of eligible babies did not receive kangaroo mother care. We did not evaluate the factors that could have been associated with improved survival (such as duration and number of days receiving kangaroo mother care sessions, breastfeeding, degree of prematurity and birth weight). We did not have a control group, and we could only compare our data with historical norms. Kangaroo mother care facilities were established across different geographical locations; however, there was no stratification or sampling to ensure systematic representation of Bangladesh's population.
Other factors could also limit applicability of the findings to other countries. Although there is a WHO standardized definition of kangaroo mother care, 10 the practice and implementation of the components of kangaroo mother care (skin-to-skin care, exclusive breastfeeding, early discharge home with appropriate follow-up and monitoring) vary widely. 17 The programme was implemented in Bangladesh with support from international and national partners with a long history of working in public health sectors. There were other contextual variables such as support from the government, involvement of health professionals and community engagement that may not be replicated in other countries.
We recognize that there are several areas of the programme that can be improved. Nonfunctioning kangaroo mother care facilities can become functioning through supportive measures such as direct visits to facilities, identifying problems and monitoring and supervision. More eligible babies should receive kangaroo mother care. In Bangladesh around 50% of babies are delivered at home and only 53% are attended by medical personnel. 14 We believe there is an opportunity to initiate kangaroo mother care at home through preparation of mothers during the antenatal period and home visits by trained health-care professionals. As the kangaroo mother care programme expands, private hospitals, where a significant number of babies are born, should be included in the programme. Further capacity-building and training Box 1. Lessons learnt for implementation of kangaroo mother care services Monitoring poor performance Nonfunctional kangaroo mother care centres need to be identified and transformed into functional centres. Improvements can be made by direct visits to health-care facilities, identifying their problems and supportive monitoring and supervision.

Capacity-building
An adequate number of health-care providers providing kangaroo mother care service from a single centre need to be trained so that the programme can continue in case of high turnover of trained staff, especially nurses.

Empowering midwives and community health-care providers
In Bangladesh, around half of deliveries that take place at home are followed up by community health-care providers. 14 On the other hand, normal deliveries at government hospitals are mainly supported by midwives. Empowering these two groups with kangaroo mother care services can improve programme outcomes.

Preventing prematurity and low birth weight
The kangaroo mother care programme should run in conjunction with antenatal services to prevent premature birth and low birth weight through interventions such as nutritional management, frequent antenatal follow-up and health education.

Expanding the data set
We intend to collect more variables on kangaroo mother care (such as duration of kangaroo mother care, duration of each session, incidence of hypothermia and infection, percentage of babies requiring oxygen, and babies on nasogastric or orogastric tube feeding during kangaroo mother care). A separate data set can be created for babies born in the community and private hospitals.

Case tracking
Most of the mothers who undertook kangaroo mother care in Bangladesh were lost to follow-up. Lack of motivation and limitations in contacting parents were the major challenges. Therefore, family participation during kangaroo mother care counselling and using a designated mobile phone in the hospital for tracing mothers could be helpful.

Recognizing staff performance
Kangaroo mother care events could be arranged in communities so that facilities can share information about their care practices and outcomes and can be rewarded for good performance. The best performing nurse or midwife from each facility could be given a public recognition award to acknowledge their contributions and motivate others.

Boosting research
Research on kangaroo mother care should be undertaken in different hospitals to determine the best ways to overcome challenges. Similarly, existing data sets such as DHIS2 can also be used for future large-scale research.
Kangaroo mother care, Bangladesh ANM Ehtesham Kabir et al.
will provide adequate manpower to keep kangaroo mother care facilities running in case of staff resignations or relocation. Incentives can be provided to the betterperforming kangaroo mother care facilities and events can be organized to showcase innovations within facilities. We also noted poor post-discharge follow-up visits, especially the second, third and fourth follow-ups. Probable factors contributing to poor follow-up were loss of contact with mothers, lack of appropriate counselling to mothers and low awareness of the programme among recipients. We have been rolling out mitigation measures progressively such as increasing the number of trained personnel, increasing community involvement with social and behavioural change communication, procurement of essential commodities for the service, and education to further improve postdischarge follow-up. Finally, we intend to expand our database to include relevant variables for further research and more effective evaluation. A separate data set can be created for babies born in the community and in private hospitals.
We were able to develop a set of recommendations for scaling up and strengthening of the programme (Box 1). We have learnt that strong commitment from the government and development partners and a positive working environment are important for a kangaroo mother care programme to run successfully in low-resource settings. However, such programmes must have continuous monitoring, adjustments and improvements. For example, at the initial phase of our programme, the number of babies receiving kangaroo mother care was low. The enrolment reached a low of just over 1% soon after the programme started. We explored factors contributing to low enrolment through informal surveys among the service providers. We determined lack of knowledge about kangaroo mother care resulting in lack of motivation from the community and service providers and limited numbers of trained staff. As a direct consequence of these factors there were many nonfunctional kangaroo mother care corners that were major contributors to poor enrolment.
Encouragingly, stringent continuous monitoring helped us to detect this trend and allowed us to implement corrective measures. These measures included training additional health-care providers and implementing internal and external supervision of kangaroo mother care activities, with resultant improvement from April 2018. The Directorate General of Health Services and Directorate General of Family Planning, both under the health ministry, plan to train an additional 41 700 and 20 820 health professionals on kangaroo mother care, respectively, for the period 2020-2022. In addition, the Directorate General of Health Services is planning to expand kangaroo mother care corners in another 125 facilities by 2020-2021, covering all district-and subdistrictlevel health facilities by 2022. Our preliminary data provide useful lessons for implementation of kangaroo mother care services at the national level