Helping small babies survive: an evaluation of facility-based Kangaroo Mother Care implementation progress in Uganda

Introduction Prematurity is the leading cause of newborn death in Uganda, accounting for 38% of the nation's 39,000 annual newborn deaths. Kangaroo mother care is a high-impact; cost-effective intervention that has been prioritized in policy in Uganda but implementation has been limited. Methods A standardised, cross-sectional, mixed-method evaluation design was used, employing semi-structured key-informant interviews and observations in 11 health care facilities implementing kangaroo mother care in Uganda. Results The facilities visited scored between 8.28 and 21.72 out of the possible 30 points with a median score of 14.71. Two of the 3 highest scoring hospitals were private, not-for-profit hospitals whereas the second highest scoring hospital was a central teaching hospital. Facilities with KMC services are not equally distributed throughout the country. Only 4 regions (Central 1, Central 2, East-Central and Southwest) plus the City of Kampala were identified as having facilities providing KMC services. Conclusion KMC services are not instituted with consistent levels of quality and are often dependent on private partner support. With increasing attention globally and in country, Uganda is in a unique position to accelerate access to and quality of health services for small babies across the country.


Introduction
Prematurity is the leading cause of newborn death in Uganda, accounting for 38% of the nation's 39,000 annual newborn deaths [1]. After pneumonia and malaria it is the third leading cause of deaths amongst children under age five [1]. Over 200,000 or 14% of Ugandan babies are born prematurely (before 37 weeks gestation) [2]. Those who survive may face a lifetime of disability with limited access to supportive services. These figures illustrate the urgency of addressing this burden and echo what is seen as an emerging priority in global public health [3]. Approaches to improve quality of care of preterm infants in health facilities were ranked second out of 82 questions in global research priority setting for preterm babies [4].
One of the highest impact interventions for newborn survival and health is kangaroo mother care (KMC) [5]. It is a low-tech and costeffective intervention in which mothers serve as human "incubators" for their newborns. KMC comprises a set of care practices for low birth weight newborns -including continuous skin-to-skin contact, establishing breastfeeding, and close follow up after discharge from a health facility [6]. KMC has been shown to reduce neonatal mortality by over 50% amongst babies weighing less than 2000g at birth [7]. It has also been found to be highly effective in reducing severe morbidity, particularly from infection [8]. Other effects when compared with incubator care include the reduction of hypothermia, severe illness, lower respiratory tract disease, and length of hospital stay. Babies cared for with KMC show improved weight gain, length, and head circumference, breastfeeding, and mother-infant bonding [8,9]. Despite convincing evidence, KMC uptake is low and only a very small proportion of newborns who could benefit from KMC receive it [10].
Compared to some other countries in the region, KMC was highlighted in Uganda at a late stage. KMC was first introduced in Mulago Hospital in 2001, but there was very little further spread of the practice beyond this national level teaching hospital. KMC remained "under the radar" while a more comprehensive approach to policy change encompassing a broader package of facility and community interventions for newborn survival was adopted nationally [11].
One of the events in Uganda that brought KMC into the public domain was a provocative editorial cartoon and newspaper article from August 7, 2007 entitled "Government tells mothers to use charcoal stoves as makeshift incubators" [12]. According to this article, the Director-General of Health Services recommended the use of the sigiri, a charcoal stove, to keep premature babies warm in poor, rural areas. This was followed by a period of advocacy for more appropriate methods of thermal care. On August 29, 2007 articles on KMC were published in both national daily newspapers in which the KMC method and its advantages were explained [13,14].
According to a newborn situation analysis commission by the Ministry of Health and overseen by the country's national newborn steering committee, a number of major challenges for newborn health were identified and included the limited availability of special services such as KMC for the care of preterm babies at health centre level, "inadequate knowledge of newborn care among health providers, a lack of institutional support for evidence-based low-cost interventions, such as KMC, and a critical lack of trained staff" [15].
Hence preterm babies born at home or at lower levels of care were almost always referred to hospitals and if referral was not possible, lanterns and coal stoves were used to provide extra heat in the rooms. Even in the hospitals, locally made incubators were used but they were prone to breakdown and suboptimal functioning due to irregular power supply [15]. As a result, the national newborn steering committee recommended immediate action at health facility level to increase the speed of roll-out of KMC in facilities starting at the health-centre level IV and above, with strong links to community follow up [15].
An evaluation of KMC services in Uganda was undertaken in order to gauge the progress towards scaling up KMC following these recommendations. The evaluation aimed at systematically measuring the scope and institutionalisation of KMC services and to describe barriers and facilitators to sustainable KMC services.

Methods
A cross-sectional, mixed-method evaluation design was used to analyse the country's progress with KMC implementation against a previously developed stages-of-change model [16] which has been used elsewhere [17,18]. Approval to conduct the evaluation was Facilities were assessed by two progress monitors by means of standardized, key-informant interviews and an observation inventory covering the following aspects of service and types of practices: the health care facility (including its baby-friendly status); neonatal and KMC facilities; skin-to-skin practices; history of KMC implementation; involvement of internal role players; physical and financial resources; KMC space (continuous and intermittent KMC); feeding and weight monitoring; referral, discharge and follow up; record keeping and documentation; KMC education; staffing issues (orientation and training; rotations); strengths and challenges [16].
Each facility received a score out of a total of 30. The scoring is divided according to six stages of institutionalisation, with each stage having a weighted score: create awareness (2 points); commit to implement (2 points); prepare to implement (6 points); implement (7 points); integrate into routine practice (7 points); sustain practice (6 points) [16]. Though KMC does not require much in terms of materials, this was raised as a concern. For the most part mothers provided their own materials to tie the baby in the KMC position. In one facility, linen was provided but not laundry service. Private, external support  Table 2). An informant from one hospital described the absence of internal support, "So far I've only seen Save (the Children)."

Results
In all facilities there was staff that had been trained in KMC, either as a stand-alone training or as part of a broader training in essential newborn care. According to some informants, the introduction to KMC was brief but it resonated: "Before, I had no idea KMC could save a baby." A total of 262 health workers were reportedly trained in KMC both on and off-site. However within the last 3 years, only 22% of the staff reportedly trained in KMC were still working with newborns at the time of the evaluation, due to attrition and staff rotation. Three hospitals could produce a written feeding policy, whereas 6 hospitals had a feeding job aid for calculating the volumes of feeds that was displayed on the wall. There was written evidence of expressed breastmilk feeding in only 3 facilities.

KMC practice and documentation
Babies' response to KMC and feeding was also monitored and in 7 facilities it was indicated that they weighed the babies regularly.
Four facilities weighed once per day, 2 on alternate days and 1 weekly. Only one hospital did not have a scale while 3 of them had a mechanical one. Change in weight of babies was benchmarked on admission weight and discharge weight was also taken and recorded in varying types of documents including nursing and doctors' notes, the baby's file (e.g. observation charts), the mother's chart, the KMC register and the discharge form. There was evidence of record keeping in all health facilities including locally-adapted KMC registers. In 4 hospitals evidence was found of doctors' daily notes, which could include a prescription for the commencement of KMC.
Two facilities recorded KMC on the discharge letter/form and one in the baby's health booklet.

Facility discharge and follow-up care
Decision to discharge a baby from the health facility was a joint effort between doctors and nurses in 8 of the health facilities while in 3 of the facilities it was solely made by nurses. There were differences in the reported criteria for discharge followed and documentation was lacking. Only 2 facilities had discharge checklists or procedures.
Four facilities had evidence of a good follow-up system and could provide records of visits. Two could provide some evidence of follow up, whereas in the 2 health centres and 3 district hospitals no evidence could be provided. Babies were followed up in either the neonatal unit/KMC space (n=4), or the maternity ward (n=3), or the Page number not for citation purposes 5 paediatric outpatient clinic (n=4). Only one hospital had a special premature clinic on Fridays. Keeping records for follow up was found to be of "good" quality in one facility whereas it was "average" for 6 facilities and deemed "poor" in 4.
Follow up at the facility where the baby had been born or had received KMC was reported to be done until the baby reached a specific weight or a specific age. Weights mentioned were 2 kg (n=1), 2.5 kg (n=5) and 3 kg (n=2). However, estimates by informants on the percentage of babies returning for follow up varied between "few" in one health centre, 30 to 50% in 3 other facilities, 80 to 90% in 2 hospitals and 95% or more in 2 of the private, not-for-profit hospitals. The main reason given for poor follow-up rates was distance from the facility. limited given the breadth of material both of these packages cover through in-service training. This is consistent study findings where informants indicated that KMC introduction was limited and brief [17,18].

Discussion
All of the facilities known to be practicing KMC were located in Kampala, the two Central regions, Southwest and the East-Central region. While this has been advantageous for teaching and learning

Competing interests
The authors declare no competing interest.  Table 1: facility scores and interpretation of the scores Table 2: overview of support with equipment and materials