Kangaroo mother care: EN-BIRTH multi-country validation study

N/A

Background: Kangaroo mother care (KMC) reduces mortality among stable neonates ≤2000 g. Lack of data tracking coverage and quality of KMC in both surveys and routine information systems impedes scale-up. This paper evaluates KMC measurement as part of the Every Newborn Birth Indicators Research Tracking in Hospitals (EN-BIRTH) study. Methods: The EN-BIRTH observational mixed-methods study was conducted in five hospitals in Bangladesh, Nepal and Tanzania from 2017 to 2018. Clinical observers collected time-stamped data as gold standard for mother-baby pairs in KMC wards/corners. To assess accuracy, we compared routine register-recorded and women’s exit survey-reported coverage to observed data, using different recommended denominator options (≤2000 g and ≤ 2499 g). We analysed gaps in quality of provision and experience of KMC. In the Tanzanian hospitals, we assessed daily skin-to-skin duration/dose and feeding frequency. Qualitative data were collected from health workers and data collectors regarding barriers and enablers to routine register design, filling and use. Results: Among 840 mother-baby pairs, compared to observed 100% coverage, both exit-survey reported (99.9%) and register-recorded coverage (92.9%) were highly valid measures with high sensitivity. KMC specific registers outperformed general registers. Enablers to register recording included perceptions of data usefulness, while barriers included duplication of data elements and overburdened health workers. Gaps in KMC quality were identified for position components including wearing a hat. In Temeke Tanzania, 10.6% of babies received daily KMC skin-to-skin duration/dose of ≥20 h and a further 75.3% received 12–19 h. Regular feeding ≥8 times/day was observed for 36.5% babies in Temeke Tanzania and 14.6% in Muhimbili Tanzania. Cup-feeding was the predominant assisted feeding method. Family support during admission was variable, grandmothers co-provided KMC more often in Bangladesh. No facility arrangements for other family members were reported by 45% of women at exit survey. Conclusions: Routine hospital KMC register data have potential to track coverage from hospital KMC wards/corners. Women accurately reported KMC at exit survey and evaluation for population-based surveys could be considered. Measurement of content, quality and experience of KMC need consensus on definitions. Prioritising further KMC measurement research is important so that high quality data can be used to accelerate scale-up of high impact care for the most vulnerable.

The EN-BIRTH study was a mixed-methods observational study comparing data from clinical observers (considered the gold standard) to women’s exit survey-reported and register-recorded coverage (Fig. 1). Detailed information regarding the research protocol, methods and analysis have been published separately [18, 25]. Data were collected between June 2017–July 2018 in five public comprehensive emergency obstetric and newborn care (CEmONC) hospitals in three high mortality burden countries: Maternal and Child Health Training Institute (MCHTI), Azimpur and Kushtia General Hospital in Bangladesh (BD); Pokhara Academy Health Sciences in Nepal (NP); Temeke Regional Hospital and Muhimbili National Referral Hospital in Tanzania (TZ). Study participants for this analysis were consenting women with babies receiving routine KMC after admission to KMC wards/corners including inborn babies (born in the study hospitals) and outborns (born elsewhere). Stata version 14 was used for all quantitative analyses [26]. Results are reported in accordance with STROBE statement checklists for cross-sectional studies (Additional file 1). KMC validation design, EN-BIRTH study Research clinical observers worked in shifts covering 24 h per day. Observation was performed without interacting with the mother-baby pair. Time-stamped observation data were collected on components of KMC care. The observer collected the initial observation data as soon as possible after admission to KMC ward/corner. Admission weight was collected from individual case notes. Regular follow-up point observations for KMC position, and feeding were hourly in KMC wards in Tanzania and every 12 h in KMC corners in Bangladesh and Nepal. Women were interviewed after discharge before exit from hospital with close-ended questions regarding KMC. Researchers extracted individual mother-baby KMC data from routine hospital registers. Register designs were described and summarised. Data were collected using a custom-built android tablet-based app developed in such a way that interviewer and register extractor data collectors could not access clinical observation data, however, data were linked at individual level. Metadata for observation, survey and register are shown in Additional file 2. Definitions of KMC coverage during admission to the KMC ward/corner are shown in Table 1. To assess accuracy at population-level (in hospital), we independently calculated and compared observed, exit survey-reported and register-recorded KMC coverage for all admitted mother-baby pairs admitted to KMC ward/corner (Fig. ​(Fig.1).1). Individual-level validity “diagnostic test” methods were calculated using two-way tables, excluding missing pairwise data. Where column totals were ≥ 10 counts, we calculated sensitivity, specificity, negative predictive value, positive predictive value, area under the curve, and inflation factor; otherwise we present percent agreement [27]. All calculations were stratified by hospital and with 95% confidence intervals (assuming a binomial distribution and using Stata’s proportion and metaprop commands). We calculated I2 and τ2 to assess heterogeneity between hospitals and combined hospital-specific results using random effects meta-analysis approach. Definition of terms for KMC sample and measurement, EN-BIRTH study 1. Upright (vertical) position 2. Skin-to-skin – newborn with caregiver’s chest 3. Legs flexed in a ‘frog position’ 4. Cheek of newborn in contact with caregiver’s chest 5. Fixed firmly to caregiver’s chest (with cloth or wrap) Caregiver – mother or other family member Arrangement Pre-discharge counselling (A) (B) (C) (D) (E) (F) refer to columns in Fig. ​Fig.44 To determine reliability of the observational data, we calculated inter-rater Cohen’s Kappa coefficients for the same 5% sample observed by both supervisors and data collectors. We also calculated Kappa coefficients for a 5% sample of double-extracted study register data. We explored KMC coverage measurement using two possible newborn admission weight denominator options: 1) ≤2000 g as the true denominator for ‘newborns in need of KMC’ as recommended by WHO, 2) ≤2499 g as some national programmes recommend KMC for all low birthweight (LBW) babies. We used KMC ward/corner admission weight as outborns may not be weighed at birth and inborns may be transferred after stabilisation for days/weeks on other neonatal wards. We measured coverage of key recommended components of KMC as markers of high-quality content KMC, to determine how coverage gaps vary depending on the measure used. We designed a gap analysis figure for (A) total eligible population of newborns admitted to KMC. Among those receiving any KMC (upright/vertical and/or skin-to-skin) (B), the KMC components used as markers of high quality KMC or “right” position content evaluated were: (C) wearing a hat, (D) five position components: 1. Upright/vertical 2. Skin-to-skin contact on caregiver’s chest 3. Legs flexed in a ‘frog position’ 4. Cheek in contact with caregiver’s chest 5. Fixed with cloth/wrap to caregiver’s chest. We further selected the subset of KMC baby days with sufficient point observations in each 24 h period to capture KMC quality for: daily duration (hereafter called KMC daily dose) ≥20 position point observations and ≥ 8 feeding observations. We calculated: (E) KMC skin-to-skin daily dose ≥20 h/day (assuming each point observation was a proxy for 1 hour of KMC), 12–19 h and < 12 h/day [5] (F) regular feeding ≥8 times/day. To assess components of quality of experience of care, at each point observation we calculated the proportion of KMC given by the mother alone or by another family member’s help. We asked women to report reasons for not doing KMC, grouping them as mother-related and baby-related. At exit-survey, we asked whether there were practical arrangements for family members to be involved during KMC admission and if pre-discharge counselling had been received. We evaluated KMC register documentation issues as part of the wider barriers and enablers objective in the EN-BIRTH study. Two tools were designed: a) semi-structured in-depth interview (IDI) guide and b) semi-structured focus group discussion (FGD) guide, both informed by the Performance of Routine Information System Management (PRISM) conceptual framework [28]. We interviewed two purposively sampled groups of respondents: hospital health workers involved in KMC register recording and EN-BIRTH study data collectors, sampling until saturation was reached. Qualitative data were coded using pre-identified codes based on PRISM using NVivo 12 for data management. Our analysis was based on applying the same methodology as an associated EN-BIRTH paper exploring barriers and enablers to routine labour ward register recording [18]. We identified emerging themes for KMC register recording across all five hospitals by the three register process categories 1) Design 2) Filling and 3) Perceived utility.

• Dima Health