Association between person-centered maternity care and newborn complications in Kenya

Objective Despite the recognized importance of person-centered care, very little information exists on how person-centered maternity care (PCMC) impacts newborn health. Methods Baseline and follow-up data were collected from women who delivered in government health facilities in Nairobi and Kiambu counties in Kenya between August 2016 and February 2017. The final analytic sample included 413 respondents who completed the baseline survey and at least one follow-up survey at 2, 6, 8, and/or 10 weeks. Data were analyzed using descriptive, bivariate, and multivariate statistics. Logistic regression was used to assess the relationship between PCMC scores and outcomes of interest. Results In multivariate analyses, women with high PCMC scores were significantly less likely to report newborn complications than women with low PCMC scores (adjusted odds ratio [aOR] 0.39, 95% confidence interval [CI] 0.16–0.98). Women reporting high PCMC scores also had significantly higher odds of reporting a willingness to return to the facility for their next delivery than women with low PCMC score (aOR 12.72, 95% CI 2.26–71.63). The domains of Respect/Dignity and Supportive Care were associated with fewer newborn complications and willingness to return to a facility. Conclusion PCMC could improve not just the experience of the mother during childbirth, but also the health of her newborn and future health-seeking behavior.

A recent systematic review on interventions to improve PCMC, such as continuity of midwifery care, decision-making tools, and information provision on various outcomes (including maternity care, perinatal health, and mental health outcomes) found mixed results regarding the impact of PCMC interventions on labor and delivery outcomes. 5 There was also poor evidence for the impact of these interventions on perinatal-related outcomes. 5 Outside of maternity care, however, person-centered care approaches in primary care settings are associated with reduced specialty care clinic visits, hospitalizations, and healthcare costs. 6 More evidence is therefore needed on how PCMC impacts health outcomes, including newborn outcomes.
Quality of care is particularly relevant in Kenya, where 90% of maternal deaths at major referral hospitals in 2017 resulted from substandard care. 6 Studies have also documented poor PCMC in Kenya. One study found that on a scale of 0-100, PCMC was less than 70% in Kenya. 7 In addition, one in five women reported feeling humiliated during labor and delivery. 8  The primary aim of the present study was to explore the association between PCMC and newborn-related outcomes in Kenya, including newborn complications and rates of immunization. A secondary aim was to examine the association between PCMC and a woman's intention to deliver in the same facility in the future. It was hypothesized that women who experience higher levels of PCMC will report lower levels of newborn complications, higher newborn immunizations, and greater willingness to deliver in the same facility for her next delivery.

| Study participants and recruitment
Respondents were recruited from seven government health facilities in Nairobi and Kiambu counties in Kenya. Baseline and follow-up surveys were completed between August 2016 and February 2017. Participants were recruited during their recovery in the post-partum ward and were eligible to participate if they were aged 15-49 years, had a normal delivery (i.e. not scheduled cesarean delivery) within the last 7 days, possessed a text-equipped phone, and reported feeling well enough to participate. In total, 531 respondents completed the baseline survey, while 62 refused participation and 124 were ineligible due to not having their own textequipped phone or not delivering in the past 2 weeks. Written informed consent was obtained before conducting study procedures.
Baseline respondents received airtime credit worth approximately US$1.00 for their participation. Follow-up surveys took approximately 10 minutes and were conducted using mSurvey, in which questions were sent via phone and respondents answered through free text messages. Respondents received follow-up surveys at 2, 6, 8, and F I G U R E 1 Flow chart of baseline and follow-up sample sizes and response rates.

| Survey measures
Dependent variables: The primary outcomes of interest were newborn complications, visiting a health facility for newborn immunizations, and women's willingness to deliver in the same facility for her next delivery. Newborn complications were assessed at a 2-week follow-up by asking women to report whether their newborn had experienced any health complications, including jaundice, fever, difficulty breathing, or other problems. Having visited a health facility for the newborn's immunizations was assessed at the 6-week and 10week follow-ups; indicating "Yes" at either follow-up was defined as a newborn having received immunizations in the newborn period. The mother's willingness to return to the same facility for her next delivery was assessed at the 8-week follow-up. PCMC scores, as well as each sub-scale score, as "low," "medium," or "high," with scores in the approximate lower 25th percentile defined as "low" and scores in the top 75th percentile defined as "high." Values are given as number (percentage) or mean (standard deviation). Percentages may not add up to 100 due to rounding.
T A B L E 1 (Continued) T A B L E 1 Sociodemographic characteristics and postpartum quality of care received during delivery (n=413). total scores as either "low," "medium," or "high" postpartum quality of care received using the same percentile cutoffs as PCMC above.

Characteristic N (%) or mean (SD) a
In addition, we included information on sociodemographic characteristics, including current age, number of births, marital status, educational attainment, employment status, whether they were born in Nairobi or Kiambu County (as an indicator of internal migration), and whether they were covered under a health scheme or health insurance, as well as self-rated health status and experience of maternal complications during labor or delivery.

| Analyses
Data were analyzed with descriptive, bivariate, and multivariate statistics using StataSE version 15. 9 Bivariate and multivariate analyses were completed using women who had data at baseline and completed at least one follow-up survey (i.e. had valid data for at least one outcome of interest). A total of 413 women had complete data at baseline and at least one follow-up survey. Figure 1 illustrates baseline and follow-up rates and the number of respondents with complete data on newborn complications, newborn's immunizations, and willingness to return to the facility. Nearly 60% (n=244, 59%) of respondents had complete information on all outcomes of interest (i.e. completed all follow-up surveys).
Logistic regression was used to assess the relationship between PCMC score (measured at baseline) and each of the dependent variables of interest (measured at follow-up). Logistic regression analyses were also used to assess the association between each PCMC subscale and each of the dependent variables of interest, respectively. All multivariate models controlled for sociodemographic characteristics,

| RESULTS
Sociodemographic characteristics of the study sample (n=413) are presented in Table 1. Over one-third (31.1%) of women reported their newborn experienced complications within 2 weeks of delivery (

| DISCUSSION
To our knowledge, this is the first study to assess the impact of PCMC on newborn outcomes using a validated scale for person-centered care. Although PCMC should be considered a right, and therefore provided whether or not it improves outcomes, the finding that receipt of high PCMC is associated with neonatal health will advance advocacy efforts to improve PCMC.
Perhaps the most intriguing finding in this study was that PCMC was While PCMC is associated with improvements in newborn health and women's intentions for future delivery, there is no significant association with newborn immunizations. Other factors beyond women's experiences, such as access to care or geographical proximity to the facility, may be more important for returning to a health facility for immunizations. 11 Additionally, immunizations are likely to take place in a different location from maternal care, hence they may not be influenced by the experience in the birthing facilities.
There are several limitations to this study. First, newborn complications are self-reported by women and it is unclear whether women can accurately recognize newborn complications which might result in misreporting. Second, as women were asked to report PCMC while in the facility, social desirability bias may be present. One study found that women underreported mistreatment in a facility compared to direct observation. 12 Third, because only women with complete cases at baseline and follow-up were included, the sample in the present study was limited in terms of the capability of the statistical power to detect differences. Loss to follow-up is another limitation: only 46% of women from baseline responded to all follow-up surveys. In assessing differences between those with and without valid data at each follow-up using Pearson χ 2 tests, significant differences were detected in marital status and education at week 2; however, no significant differences in sociodemographic characteristics or PCMC score were detected at weeks 6, 8, and 10. This suggests that estimates are not highly biased by missing data. Lastly, the eligibility criteria of needing a text-equipped phone may have slightly biased the sample in the present study towards wealthier, literate women. person-centered care strategies in their facilities. Future research should assess the impact of PCMC on other health outcomes, including maternal complications, breastfeeding practices, and follow-up visits. Every woman should be treated with dignity and respect from a rights-based perspective. These findings will, however, facilitate PCMC advocacy from a health benefits perspective: that PCMC could improve not just the experience of women but also the health of her newborn.

AUTHOR CONTRIBUTIONS
MS conceived of the study, oversaw data analyses, and drafted the manuscript. AL analyzed the data and drafted parts of the manuscript. PA, NDS, and GG wrote parts of the manuscript and contributed to interpretation of the data. All authors have read and approved the final manuscript.
T A B L E 5 Multivariate logistic regressions examining the relationship between PCMC sub-scales and newborn and health-seeking behaviorrelated outcomes. a PCMC sub-scale score b Newborn experienced complications within 2 wk of delivery (n=313)

Visited facility for newborn's immunizations (n=369)
Willing to return provider/facility for next delivery (n=346)