BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 29, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

CAN A MIDWIFERY-LED CONTINUITY MODEL IMPROVE MATERNAL SERVICES IN A LOW RESOURCE SETTING? – A NON-RANDOMIZED CLUSTER INTERVENTION STUDY IN PALESTINE

ForJournal: peerBMJ Open review only

Manuscript ID bmjopen-2017-019568

Article Type: Research

Date Submitted by the Author: 12-Sep-2017

Complete List of Authors: Mortensen, Berit; Oslo Universitetssykehus Intervensjonssenteret; Universitetet i Oslo Institutt for klinisk medisin Lukasse, Mirjam; Oslo and Akershus University College, Norway, Faculty of Health Sciences; Norwegian University of Science and Technology, Department of Obstetrics and Gynaecology Diep, Lien; University of Oslo, International Health; Oslo University Hospital, Unit for biostatistics and epidemiology Lieng, Marit; Oslo University Hospital, Ullevål, Department of Obstetrics Awad, Amal; Palestinian Ministry of Health Suleiman, Munjid; Palestinian Ministry of Education and Higher Education Fosse, Erik; The Intervention Centre, Oslo University Hospital

Rikshospitalet, Oslo, Norway; Institute for clinical medicine, Faculty of http://bmjopen.bmj.com/ Medicine, University of Oslo, Oslo, Norway

Primary Subject Obstetrics and gynaecology Heading:

Secondary Subject Heading: Global health, Public health, Health services research, Mental health

Continuity of care, Midwifery, Organisation of health services < HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Quality in health care <

Keywords: on September 29, 2021 by guest. Protected copyright. HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Community child health < PAEDIATRICS, Maternal care

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 CAN A MIDWIFERY-LED CONTINUITY MODEL IMPROVE MATERNAL 4 SERVICES IN A LOW RESOURCE SETTING? – A NON-RANDOMIZED CLUSTER 5 6 INTERVENTION STUDY IN PALESTINE 7 1 ,2 3,4 5 2,6 8 Berit Mortensen MCs , Mirjam Lukasse PhD , Lien My Diep MSc , Marit Lieng PhD , 9 Amal Abu Awad, PhD7, Munjid Suleiman MCs8 Erik Fosse PhD1,2 10 11 12 Corresponding author: Berit Mortensen, Wilhelm Færdens vei 4b, 0361 Oslo, 13 14 [email protected] mobile: 0047-93266113 15 Abstract 16 For peer review only 17 Objectives To improve maternal health services in rural areas the Palestinian Ministry of 18 Health launched a midwifery-led continuity model in the West Bank in 2013 in cooperation 19 20 with Norwegian Aid Committee. Midwives were weekly deployed from governmental 21 22 hospitals to provide ante- and postnatal-care in rural clinics. We studied the intervention`s 23 impact on utilization and quality indicators of maternal services after two years’ experience. 24 25 Design: A non-randomized, intervention design was chosen. The study was based on registry 26 27 data only available at cluster level, two years before (2011 & 2012) and two ears after (2014 28 & 2015) the intervention. 29 30 Setting: A total of 53 primary health care clinics in Nablus and Jericho regions were stratified 31 for inclusion. 32 http://bmjopen.bmj.com/ 33 Primary and secondary outcomes: Primary outcome was number of antenatal visits. 34 35 Important secondary outcomes were number of referrals to specialist care and number of 36 postnatal home-visits. Differences in changes within the two groups before and after the 37 38 intervention were compared by using mixed effect models. 39 Results: 14 intervention-clinics, and 25 control-clinics were included in the study. Number of 40 on September 29, 2021 by guest. Protected copyright. 41 antenatal visits increased by 1.16 per woman in the intervention-clinics while it declined by 42 43 0.39 in the control-clinics, giving a significant difference in change of 1.55 visits [95% CI 44 0.90 – 2.21]. A significant difference rate in number of referrals was observed between the 45 46 groups 3.65 [2.78 – 4.78] as number of referrals increased by a rate ratio of 3.87 in the 47 48 intervention group while in the control the change-ratio was only 1.06. 49 Home-visits increased substantially in the intervention group, but decreased in control group 50 51 giving a difference RR 97.65 (45.20 - 210.96) 52 Conclusion: The Palestinian midwifery-led continuity model improved utilization and some 53 54 quality indicators of maternal services. More research should be done to investigate if the 55 56 model had effect on individual health outcomes and satisfaction of care. 57 58 1 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 Trial registration: ClinicalTrials.gov Identifier: NCT03145571 4 5 6 Strengths and Limitations of the study 7 • The pragmatic approach strengthens the applicability to real life settings. 8 9 • The high number of clusters and the robust cluster data strengthen the study. 10 11 • A randomized allocation of clusters was not possible because the implementation of 12 the midwifery-led continuity model started before the study was planned. 13 14 • The ministry implemented the program in the clinics they found appropriate which 15 16 could leadFor to bias. peer review only 17 • The facility-based registry did not include data on individual level. 18 19 INTRODUCTION 20 1 21 As a low-middle-income country under occupation, Palestine depend largely on foreign aid. 22 The Palestinian Authority is responsible for the health services for the Palestinian people 23 24 within the occupied territories of West Bank and Gaza. In 2013, the Palestinian Ministry of 25 Health, registered 61 405 births and a fertility rate of 4.0 per woman on the West Bank.2 26 27 Maternal health services were provided by the Palestinian government and private, and non- 28 29 governmental organisations. Less than 1% of the women gave birth at home. Governmental 30 facilities covered 45.6% of antenatal-care in 2013. The Palestinian Multiple Indicator Cluster 31 32 Survey from 2014 found that 66.4% of rural women gave birth in governmental hospitals. Of http://bmjopen.bmj.com/ 33 34 them 40.7% left hospital within 6 hours postpartum and 73% did not receive any additional 35 postnatal-care.3 36 37 In 2009 the Palestinian governmental maternal services were described as of poor quality, 38 overcrowded, understaffed, including short antenatal visits lacking content, and dissatisfaction 39 40 4 5

with care. Overcrowded labour rooms prohibited women bringing a birth companion. on September 29, 2021 by guest. Protected copyright. 41 42 Midwives had restricted scope of practice and little autonomy, and few worked in antenatal- 43 care.6 The clinics were not able to implement postnatal home-visits as required by 44 45 governmental standards.6 Poor women were less likely to have postnatal-care.7 46 47 Poverty, deprived infrastructure, military checkpoints and armed Israeli settlers restricts 48 freedom of movement and reduces access to central health-facilities and legal aid in rural 49 8,9 50 areas. Rural women depend more on governmental facilities than women in urban areas, as 51 private services are scarce.10 52 53 Several studies describe how midwifery-led continuity models improved health for mothers 54 11-16 55 and babies. Most studies were from high-income countries. World Health Organisation 56 (WHO) recommends implementation and research on midwifery-led continuity models to 57 58 2 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 17 3 improve quality in low- and middle-income countries. Two main ways of organizing such 4 models are described in the literature. In the case-load model one midwife cares for up to 40 5 6 women and facilitate relational continuity, while in the team-midwifery model a group of 4-6 7 8 midwives can provide care for up to 360 women through pregnancy, birth and postnatally. 9 Ideally, in both models, women are cared for during birth by a known midwife.11,17 10 11 To improve services in rural areas, the Palestinian Ministry of Health in cooperation with the 12 non-governmental humanitarian organization Norwegian Aid Committee (NORWAC), 13 14 launched implementation of a modified midwifery-led continuity, case-load-model, in 2013, 15 16 starting in NablusFor and Jericho peer Governmental review hospitals and surrounding only villages. The 17 implementation engaged several levels in the ministry of health and the involved 18 19 communities, to overcome known barriers to quality of care.18 20 21 The aim of this study was to investigate whether the Palestinian midwifery-led continuity 22 model had impact on the utilization of services and selected quality indicators at the two 23 24 regions` clinics after two years of experience. 25 26 27 METHODS 28 29 Implementation of the Palestinian Midwifery-led continuity model 30 The modified case-load model aimed at establishing relation between the pregnant woman 31 32 and a midwife during pregnancy and in the postnatal period. http://bmjopen.bmj.com/ 33 34 Assigned, trained and supportive supervised midwives from the local governmental hospital 35 provided ante-and postnatal care in governmental village-clinics and private homes, 36 37 respectively. The same midwife visited the same village weekly, and worked in the labour 38 ward the rest of the week. The midwives used designated cars to facilitate transportation. 39 40 Standard care in the clinics without this model was offered mainly from nurses or midwives on September 29, 2021 by guest. Protected copyright. 41 42 and medical doctors (GP) working only in primary health care. 43 44 45 Study design 46 47 A non-randomized intervention design was chosen to evaluate the model, based on registry 48 data at cluster level from two complete years before (2011 & 2012), and two complete years 49 50 after (2014 & 2015) the intervention. Clinics where the model was implemented were 51 compared with clinics where the model was not implemented. Both arms of the study 52 53 followed the same written governmental procedures. 54 55 56 57 58 3 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 The study was part of an implementation research project aiming at documenting the effect of 4 the midwifery-led continuity model. The study was approved by the Regional ethical 5 6 committee of South East Norway and by the Palestinian Ministry of Health. 7 8 9 Clusters 10 11 The clusters consisted of governmental primary health village-clinics in Nablus and Jericho 12 regions. There were 53 active clinics during the study period. During autumn 2013, the 13 14 midwifery model had been implemented in 16 clinics. 15 16 All clinics were Forstratified bypeer rural and urban review location, activity only period and intervention period. 17 18 19 Exclusion criteria 20 21 Clinics located less than three km from Nablus and Jericho centre were defined as urban, thus 22 pregnant women had better access to private and non-governmental services. Clinics in urban 23 24 areas were therefor excluded. Clinics opened during the study period were excluded due to 25 incomplete data. 26 27 Clinics where the intervention was prematurely terminated or introduced later, during the 28 29 study period were excluded because of contaminated data. 30 31 32 Outcomes http://bmjopen.bmj.com/ 33 34 The number of antenatal visits was chosen as primary outcome. Secondary outcomes were 35 number of pregnant women referred to higher level of care, and number of women receiving 36 37 postnatal home-visits. Other outcomes were number of women registered at the clinic for 38 antenatal-care, number of pregnant referred for abnormal blood sugar levels, number of 39 40 women seen by doctor after birth, number of newborn seen by doctor after birth and number on September 29, 2021 by guest. Protected copyright. 41 42 of total postnatal consultations for mother and newborn. 43 44 45 Statistical analysis 46 47 Aggregated data were retrieved from the governmental registry. The registry consisted of 48 anonymous recordings reported monthly from all clinics to the central statistical database in 49 50 the Ministry of Health. The registry did not include data at an individual level. 51 Mean, standard deviation and range were given for normally distributed and count variables. 52 53 Percentage and total number were given for categorical variables. 54 55 56 57 58 4 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 Change from baseline in the intervention and standard care groups, and any differences 4 between the groups changes, were examined by using mixed effects linear regression and 5 6 mixed effects Poisson regression. In the mixed models, the clinic was specified as cluster (i.e. 7 8 random variable), time and group, and interaction between time and group were treated as 9 fixed variables. 10 11 The outcome variables were divided by the number of registered pregnant women or newborn 12 to calculate average values, and were then fitted by mixed effects linear regression. Since the 13 14 computed averages are based on different numbers of pregnant women, the model must 15 16 consider the resultingFor heteroscedasticity. peer reviewThis was done by computingonly variance weights for 17 each average before analysis, and including it in a variance formula in the model. 18 19 Additionally, an offset variable was used to adjust for the total number of (individual) 20 21 registrations that were under risk in the mixed effects Poisson regression. 22 23 24 Covariates like population size, employed community midwife and laboratory equipment at 25 the clinic, check points, and distance from hospital, which may have had impact on the study 26 27 groups, were also included in the model for adjusting (table 1). 28 29 Table 1 Characteristics of clusters* 30 Characteristics Intervention Control 31 Population served mean 3402 4636 32

min/max 1000/7554 1875/11017 http://bmjopen.bmj.com/ 33 Distance to hospital (km) mean 23 12J6 34 35 min/max 5/59 3/28 Number of clinics with employed 0 8 36 community midwife 37 Number of clinics with laboratory 5 10 38 39 Additional clinics in village (NGO) 4 2 40

Number of clinic with regular military checkpoints on September 29, 2021 by guest. Protected copyright. 6 14 41 between village and hospital 42 * 14 clinics with intervention and 25 clinics with standard care(control) 43 44 45 Adjusted regression coefficient as means and rate ratios with 95% confidence intervals were 46 given. 47 48 The p-value was two-sided and difference level of 0.05 or less was considered significant. 49 50 The analyses of mixed effect models were performed with R version 3.4 and STATA version 51 14. Descriptive analyses were carried out using IBM-SPSS version 21 for Windows. 52 53 54 55 56 57 58 5 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 RESULTS 4 After stratification, 39 clinics were included in the study, 14 as intervention and 25 as control 5 6 clinics (figure 1). In total, 10 034 women booked at the 39 included clinics during the study 7 8 period, 2 784 in the intervention clinics and 7 250 in the control clinics. 9 The clinics location is presented in figure 2. The Palestinian Ministry of Health confirmed that 10 11 no other activities were introduced unequally to the groups during the study period. 12 Descriptive statistics of primary and secondary outcome variables such as number of 13 14 individuals registered, mean, standard deviation and range for the two timepoints in the 15 16 intervention and Forstandard carepeer are presented review in table 2. only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 6 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 Table 2 Descriptive statistics for primary and secondary outcomes at baseline and two years after the implementation 4 Groups/clusters (n=39)*timepoint** N*** Mean**** SD Range 5 min max Before 1094 39.1 17.6 12 95 6 Intervention 7 Number of women registered for After 1690 60.4 33.1 16 163 8 ANC during study period Before 3180 63.6 35.9 3 168 Standard care 9 After 4070 81.4 41.6 23 213 10 Total 10034

11 Before 2220 79.3 41.6 27 176 Intervention 12 Number of newborn registered during After 2470 108.3 60.7 29 217 study period 13 Before 5416 88.2 45.9 33 291 Standard care 14 After 5771 115.4 67.1 36 298 15 total 15877 16 For peer Beforereview 4015 3.7 only0.9 1.9 5.6 17 Number of antenatal visits and mean Intervention visit per women per cluster (visits After 7994 4.7 1.2 3.0 7.0 18 divided on number of registered Before 14657 4.6 1.4 2.4 8.6 19 women) Standard care After 16769 4.2 1.1 2.4 6.8 20 Before 79 7.3% 8.3% 0 36.4% 21 Number of referrals to higher level of Intervention After 456 25.6% 14.2% 2.3% 54.1% 22 care (mean % of registerd pregnant Before 427 12.0% 11.7% 0 45.5% 23 per cluster) Standard care 24 After 549 12.8% 13.2% 0 66.7% Before 12 1.7% 4.7% 0 17.7% 25 Number of registered pregnant Intervention 26 women receiving home visits after After 721 41.8% 25.2% 0 97.5% birth (mean % of registered pregnant Before 42 1.5% 4.9% 0 25.5% 27 per cluster) Standard care 28 After 22 0.7% 2.3% 0 11.5% Before 56.5% 26.6% 20.9% 109.2% 29 Intervention Coverage - ratio betweeen of number After 70.7% 21.8% 37.3% 130.6% 30 of registered newborns and registered Before 61.4% 23.0% 6.8% 118.9% 31 pregnant per cluster ) Standard care 32 After 74.3% 16.3% 46.8% 113.0% http://bmjopen.bmj.com/ 33 Before 29 2.6% 3.4% 0 10.0% Number of referrals because of Intervention 34 After 81 4.5% 7.4% 0 30.2% unnormal blood sugar (mean % of 35 Before 90 2.6% 3.5% 0 14.9% registered pregnant per cluster) Standard care 36 After 105 2.7% 3.6% 0 13.0% 37 Before 208 12.3% 20.0% 0 0.7% Number of mothers seen by doctor Intervention 38 After 461 20.1% 32.4% 0 94.0% postnatally (mean % of registered 39 Before 534 12.1% 20.5% 0 93.0% newborn per cluster) Standard care 40 After 225 4.8% 10.0% 0 57.0% on September 29, 2021 by guest. Protected copyright. 41 Before 1670 79.6% 24.5% 29.7% 118.6% Number of newborn seen by doctor Intervention 42 After 2173 91.4% 29.8% 34.8% 172.4% postnatally (mean % of registered 43 Before 4338 85.2% 26.8% 21.8% 162.2% newborn per cluster) Standard care 44 After 5082 90.4% 21.6% 47.5% 142.1% 45 Before 3902 1.8 0.3 1.3 2.2 Intervention 46 Total postnatal consultation for After 5364 2.2 0.4 1.7 3.1 47 newborn Before 9796 1.9 0.3 1.2 2.6 48 Standard care After 10875 1.9 0.2 1.5 2.4 49 Before 1830 0.9 0.4 0.2 1.7 50 Intervention Total postnatal consultations for After 3637 1.4 0.4 0.8 2.1 51 mothers of newborn Before 5073 1.0 0.5 0.2 3.1 52 Standard care After 5399 1.0 0.7 0.2 1.6 53 54 * 14 intervention clusters and 25 Standard care clusters** Two years before intervention (2011 & 2012) and two years after intervention (2014 & 2015) *** Number of 55 total individual registrations**** The mean at clusterlevel 56 57 58 7 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 Change within the intervention and standard care group as means and rate ratios (RRs) and 4 difference between the changes within the two groups, controlled for potential confounding 5 6 covariates, are presented in table 3. 7 Table 3 Change before and after intervention in both groups and multiplicative difference of changes between the groups 8 Change in groups before and after Difference in changes between groups Outcome* Group P-value 9 Adjusted***mean (95%CI) Adjusted mean (95% CI) Intervention 1.16 (0.60 - 1.72) p<0.0001 10 Mean number of antenatal visits per pregnant 1.55 (0.90 - 2.21) 0.0007 11 Control -0.39 (-0.73 - 0.05) p=0.026

12 Number of newborn mothers who registered at Intervention 18.2% (10.0 - 26.4) p<0.0001 6.6% (-3.1 - 16.4) 0.179 13 clinic during their pregnancy Control 11.6% (6.3 - 16.9) p<0.0001 14 Number of newborn seen by doctor Intervention 12.8% (-1.8 - 27.3) p=0.085 4.7% (-12.7 - 21.9) 0.599 15 postnatally Control 8.1% (-1.3 - 17.5) p=0.089 16 For peer review only 17 Total number of postnatal consultations for Intervention 0.41 (0.26 - 0.57) p<0.0001 0.33 (0.16 - 0.52) 0.0004 18 newborn Control 0.08 (-0.02 -0.18) p=0.126 19 Total number postnatal consultations for Intervention 0.64 (0.52 - 0.77) p<0.0001 20 0.60 (0.46 - 0.75) <0.0001 mothers Control 0.04 (-0.04 -0.12) p=0.321 21 22 Outcome** Group Adjusted Rate Ratio RR (95%CI) Adjusted RR (95% CI) P-value 23 Intervention 3.87 (3.04 - 4.92) p <0.0001 24 Number of referrals to higher level of care 3.64 (2.78 - 4.78) <0.0001 Control 1.06 (0.94 - 1.21) p=0.353 25

26 Number of registered pregnant who received Intervention 37.42 (21.14 - 66.22) p<0.0001 97.65 (45.20 - 210.96) <0.0001 27 postnatal home visits Control 0.38 (0.23 - 0.64) p<0.0001 28

29 Number of referrals because of unnormal Intervention 1.78 (1.16 - 2.72) p= 0.008 1.83 (1.10 - 3.05) 0.021 30 bloodsugar Control 0.97 ((0.73-1.29) p=0.846 31 Intervention 1.94 (1.65 - 2.29) p<0.0001 32 Number of mothers seen by doctor postnatally 4.87 (3.88 - 6.10) <0.0001

Control 0.40 (0.34 - 0.47) p<0.0001 http://bmjopen.bmj.com/ 33 34 35 Mixed effect linear regression was used to analyze change in and between clusters when data had normal distribution* *Mixed-effect Poisson regression was used to analyze the change in clusters when data was not 36 normal distrubuted9 ***All outputs were adjusted for potential confounders without any change in value. Covariates adjusted for were: Distance from clinic to city hospital (km), population in village, if there were 37 additional clinic in village, clinics with community midwife, clinics with laboratory and clinics with a regular checkpoint on the way to hospital. 38 39 40

on September 29, 2021 by guest. Protected copyright. 41 42 Antenatal utilization 43 There was statistically significant difference in average change in antenatal visits between the 44 45 groups by 1.55 [1.38 – 1.54], p=0.0004, mean number of visits increased by 1.16 visits with 46 47 the new model, while standard care declined with - 0.39 visits. 48 49 50 Referrals 51 A significant difference in change between the groups number of referrals to a higher level of 52 53 care was observed giving a rate difference of 3.64 [2.78 - 4.78], p<0.0001. For the 54 55 intervention group referrals increased by a rate-ratio of 3.87, while the control group only had 56 a change rate-ratio of 1.06. 57 58 8 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 5 6 7 8 Postnatal service 9 Postnatal home-visits increased substantially at the intervention-clinics, while at the control- 10 11 clinics it dropped giving a rate difference of 97.65 [45.20 – 210.96], p<0.0001. 12 Women in the intervention group increased significantly mean number of postnatal contacts 13 14 with health services, while no increase was observed at clinics with standard care, giving a 15 16 rate difference ofFor 0. 60 (95% peer CI 0.46 - 0.75) review P<0.0001. Also, only a significant increase in mean 17 number of newborn`s healthcare-contacts were observed with the new model, while not in 18 19 group with standard care giving a rate difference of 0.33 (0.16 - 0.52) p=0.0004. 20 21 22 23 24 DISCUSSION 25 In the clinics with the midwifery-led continuity intervention a significant rise in mean number 26 27 of antenatal visits per woman was observed, whereas number of visits per woman decreased 28 29 in the clinics with standard care during the same period. It is thus likely that the improved 30 utilization was a result of the intervention. WHO recommended in 2002 a minimum of four 31 32 focused antenatal visits for healthy pregnant women.19 After evaluating new evidence, the http://bmjopen.bmj.com/ 33 34 recommendation was revised in 2016 to a minimum of eight antenatal visits to reduce 35 perinatal mortality and improve women’s satisfaction.17,18,20 The women’s increased 36 37 adherence to service in clinics with midwifery-led continuity indicates that women 38 experienced improved quality of the services. Relational continuity is an important tool to 39 40 enhance communication and thus satisfaction of care.16 The association between improved on September 29, 2021 by guest. Protected copyright. 41 42 quality of care and increased utilization is supported by several previous studies and by WHO 43 recommendations to improve utilization and quality by introducing midwife-led continuity of 44 45 care.12,14,17,18 46 47 It is likely that the pregnant woman felt safe knowing that the midwife following her 48 throughout pregnancy also worked at the public hospital where she would give birth, and that 49 50 her midwife would visit her at home after birth. Due to the heavy workload, the midwives 51 could not always be on call to attend birth. Midwifery-led continuity of care in settings with 52 53 few midwife resources and heavy workload must balance the demands on the midwives. 54 55 Although women were not guaranteed that the same midwife providing them ante- and 56 postnatal care would attend their birth, their midwife`s connection to the governmental 57 58 9 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 hospital might have reduced the alien barrier to the hospital, and restored feeling of security 4 for the rural women. 5 6 A qualitative study investigated midwives’ experience from working in a similar model in 7 8 Ramallah region from 2007 to 2011. The midwives described how the model enabled them to 9 give individualized care and how the broader scope of practice and increased autonomy gave 10 21 11 them important experience and tools to help. This can explain women adherence to service, 12 because building relationship with competent, respectful and motivated midwives probably 13 14 increased their wish to return to receive more care. 15 16 For peer review only 17 In the intervention group a change in referral mechanisms was observed. The midwives 18 19 working in the new model identified significantly more risk factors causing referral to higher 20 21 level of care, than in clinics with standard care. The identification of more women with 22 abnormal blood sugar level indicates that the model led to identification of important risk 23 24 factors. The proportion of pregnant women referred to higher level of care increased to 25.6% 25 with the new model vs. 13% with standard care during the study period. The proportion in the 26 27 new model is in line with the WHO-estimate presented in guidelines from 2001, that in 28 29 general 25% of pregnant women would need additional antenatal-care due to health 30 complications before or during pregnancy.22 The village clinics had little, if any, technical 31 32 resources to investigate risk signs, so referral to higher level of care was necessary to follow http://bmjopen.bmj.com/ 33 34 up any possible complications. 35 One important quality indicator of antenatal-care is the ability to detect possible 36 37 complications and involve specialist care when necessary. Kearns et al previously 38 demonstrated that improved referral networks was a key element for improving quality in 39 40 low-resource settings.23 The process of information and referral within the system is also on September 29, 2021 by guest. Protected copyright. 41 24,25 42 highlighted by the WHO framework as one core indicator of quality of services. Some 43 countries have much higher potential health risks than others due to poverty, high fertility and 44 45 general health challenges.10,26 A Palestinian study from 2015 revealed that 26.9% of women 46 27 47 who gave birth experienced one or more morbidities. The rise in numbers of referrals after 48 introducing the new model matched the WHO and the local estimates. Thus, it seems 49 50 reasonable to suggest that the intervention improved the referral system. 51 52 53 The results furthermore showed a substantial increase in number of postnatal-care contacts for 54 55 mothers and new-borns, including home-visits. WHO recommendation for postnatal care is a 56 minimum of three postnatal contacts; and minimum one home-visit preferably during the first 57 58 10 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 28 3 week after birth. The result from the study consequently indicates that the implementation of 4 midwifery-led continuity models may contribute to reach such goal. 5 6 Filby et al described how lack of transportation hampered quality improvements in other rural 7 29 8 resource-constrained settings. The implementation in rural Palestine included a designated 9 vehicle and driving skills for the midwives, facilitation of transportation was consequently a 10 11 key in order to reach out to the villages and home-visits. 12 The fact that women receiving midwifery-led continuity were more frequently seen by a 13 14 doctor in the clinic after birth (20.1% vs. 4.8%), in addition to the midwife, indicate improved 15 16 interdisciplinary Forcooperation. peer When midwives review undertook home-visit only and discovered health 17 problems or risks, she involved the medical doctor. The findings also showed that there is a 18 19 systematic check of newborn by doctors in all the village clinics, and the home visit from the 20 21 midwife ads to this. The increase in and variation of, postnatal contacts including home-visits, 22 make it reasonable to conclude that the midwifery model improved both utilization and 23 24 quality of postnatal care at a cluster level. 25 26 27 Limitations of the study 28 29 The study was carried out after the implementation of the midwifery-led continuity model 30 started. This prevented a randomized allocation of clinics to intervention and control clusters. 31 32 The number of midwives available in the hospitals limited the number of clinics for http://bmjopen.bmj.com/ 33 34 implementation to a total of 15 in these regions. The ministry chose to implement the program 35 in the clinics they found appropriate. The baseline data indicate that the reason for choosing 36 37 these clinics were challenges in service provision, thus improvements could be easier 38 achieved and lead to bias. 39 40 Another limitation was weak data due to the facility-based registry did not include data on on September 29, 2021 by guest. Protected copyright. 41 23 42 individual level. This is a common problem in low-and middle income countries. Lack of 43 individual reproductive health registry prevented measuring individual impact and an intra- 44 45 cluster coefficient. It was not possible to identify when women registered at the clinic, who 46 47 came back for recurrent antenatal-visits or reasons for referrals. 48 49 50 Strength of the study and further recommendations 51 The high number of clusters and the robust cluster data strengthen the study. The 52 53 organizational leadership, engagement and adherence of the multidisciplinary team strengthen 54 55 the sustainably of a complex intervention and its applicability to real life settings. 56 57 58 11 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 There were high statistical significant differences for the primary outcome and important 4 secondary outcomes, also after adjusting for possible confounders. The findings make it 5 6 reasonable to conclude that the new model had effect on the utilization and some quality 7 8 indicators of the maternal service. 9 10 11 Triangulation of methods within an implementation research framework would be useful to 12 investigate the broader effect of the implementation. This is highly recommended when 13 14 introducing evidence-based interventions to improve health service delivery in real world 15 30 16 settings where contextFor is an peer important factor. reviewFurther research only should be done to investigate 17 if the model has impact on individual health outcomes for mother and newborn, and 18 19 satisfaction of care. Previous research described midwifery-led models as cost saving, also in 20 12,31 21 developing countries , nevertheless evaluation of this model`s costs would be useful. 22 23 24 Conclusion 25 The findings make it reasonable to conclude that the new model had effect on the utilization 26 27 and some quality indicators of the maternal service. 28 29 The positive change in facility-level outcomes show that clinics with the midwifery model in 30 the regions of Nablus and Jericho improved services during pregnancy and postnatally. The 31 32 findings indicate improvement of utilization and some quality indicators linked to facility- http://bmjopen.bmj.com/ 33 34 level outcomes, such as continuity, functioning referral system and postnatal home-visits. 35 The results from this study support the expansion of the model to new areas in Palestine. We 36 37 believe the model can be useful for other low- and middle-income countries to improve 38 utilization and quality of care. 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 Authors affiliation 43 1 2 44 The Intervention Centre, Oslo University Hospital, Oslo, Norway, Institute for Clinical 45 Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway, 3 Faculty of Health 46 47 Sciences, Oslo and Akershus University College, Oslo, Norway, 4 Faculty of Health and 48 5 49 Social Sciences, University College of Southeast Norway, Oslo, Norway Oslo Centre for 50 Biostatistics and Epidemiology, Oslo University Hospital, Oslo, Norway 6 Department of 51 7 52 Gynaecology, Oslo University Hospital, Oslo, Norway, Palestinian Ministry of Health, 53 8 54 Nablus. Palestine Palestinian Ministry of Education and Higher Education, Ramallah, 55 Palestine 56 57 58 12 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 Acknowledgements The authors would like to thank the Palestinian Ministry of Health for 5 6 their efforts in implementing the model, and preparing and sharing of their health registry. We 7 8 also like to thank Norwegian Aid Committee for supporting the study. Finally, we are grateful 9 to all the involved Palestinian midwives for their courageous efforts in improving care for 10 11 women in Palestine. 12

13 14 Contributors 15 16 For peer review only 17 BM was involved with the Implementation, study design, data collection, data analysis, data 18 19 interpretation and writing, MIL with study design, data interpretation and writing, LMD with 20 21 study design, data analysis and writing, MaL was involved with study design, data 22 interpretation and writing, AAA with the implementation, data interpretation and writing 23 24 MS conducted collection and systematization of data, EF involved in study design, data 25 collection, data analysis, data interpretation and writing. BM drafted the article, figures and 26 27 tables all authors have reviewed and approved the final manuscript. 28 29 Funding 30 This work was partly supported by the Research Council of Norway through the Global 31 32 Health and Vaccination Program (GLOBVAC), project number 243706 and partly by public http://bmjopen.bmj.com/ 33 34 funding through Norwegian Aid Committee. Three of the Authors were partly employed by 35 NORWAC and were involved in the implementation and interpretation of data. The analysis 36 37 was performed by a statistician at Oslo University hospital. The corresponding author had full 38 access to all the data in the study and had final responsibility for the decision to submit for 39 40 publication. on September 29, 2021 by guest. Protected copyright. 41 42 43 Competing interests None declared 44 45 46 47 Ethics approval The study was approved by the Norwegian Regional Committee for Medical 48 Health research Ethics South East (REK) id number: 2015/1235. It was also approved by the 49 50 Palestinian Ministry of Health. 51 Data sharing statement Data was obtained from the Palestinian Ministry of Health registry. 52 53 54

56 57 58 13 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 1. World Bank website Country information, accessed September 2017 4 http://data.worldbank.org/country/west-bank-and-gaza 5 2. Palestine Ministry of Health Annual report 2013. Palestine,Nablus, Palestine Ministry 6 of Health, 2013. 7 3. Statistics PCBo. Palestinian Multiple Indicator Cluster Survey 2014. Ramallah, 8 Palestine UNFPA, UNICEF, 2015. 9 4. Rahim HF, Wick L, Halileh S, et al. Maternal and child health in the occupied 10 Palestinian territory. Lancet 2009; 373(9667): 967-77. 11 5. Wick L, Mikki N, Giacaman R, Abdul-Rahim H. Childbirth in Palestine. Int J 12 Gynaecol Obstet 2005; 89(2): 174-8. 13 14 6. Abu Awad, Amal. PhD dissertation, Implementation of postnatal protocols by nurse in 15 primary health care clinics in the West Bank: University of Wisconsin-Madison 2011. 16 7. Nabaa HAA,For Hilal GA,peer Sbeih SA, review Ghandour R, Giacaman only R. Access to care for 17 women reporting postnatal complications in the occupied Palestinian territory: a cross- 18 sectional study. The Lancet 2013; 382: S2. 19 8. Reem Al-Botmeh U. A review of Palestinian legislation from a women`s right 20 perspective. UNDP-papp-research-Legislative english.pdf UNDP, 2012. 21 9. United Nations, Office for the Coordination of Humanitarian Affairs occupied 22 Palestinian territory Area C Vulnerability Profile. http://data.ochaopt.org/vpp.aspx 23 Accessed September 2017. 24 10. Graham W, Woodd S, Byass P, et al. Diversity and divergence: the dynamic burden of 25 poor maternal health. Lancet 2016; 388(10056): 2164-75. 26 27 11. Homer C, Brodie P, Leap N. Midwifery continuity of care : a practical guide. Sydney; 28 New York: Churchill Livingstone/Elsevier; 2008. 29 12. Sandall J, Soltani H, Gates S, Shennan A, Devane D. Midwife-led continuity models 30 versus other models of care for childbearing women. Cochrane Database Syst Rev 31 2016; 4: CD004667. 32 13. ten Hoope-Bender P, de Bernis L, Campbell J, et al. Improvement of maternal and http://bmjopen.bmj.com/ 33 newborn health through midwifery. Lancet 2014; 384(9949): 1226-35. 34 14. Renfrew MJ, McFadden A, Bastos MH, et al. Midwifery and quality care: findings 35 from a new evidence-informed framework for maternal and newborn care. Lancet 36 2014; 384(9948): 1129-45. 37 15. Homer CS, Friberg IK, Dias MA, et al. The projected effect of scaling up midwifery. 38 Lancet 2014; 384(9948): 1146-57. 39 16. Homer CS. Challenging midwifery care, challenging midwives and challenging the 40 on September 29, 2021 by guest. Protected copyright. 41 system. Women Birth 2006; 19(3): 79-83. 42 17. WHO. WHO recommendations on antenatal care for a positive pregnancy experience. 43 http://apps.who.int/iris/bitstream/10665/250796/1/9789241549912-eng.pdf?ua=1 44 World Health Organization; 2016. p. 152. 45 18. Nair M, Yoshida S, Lambrechts T, et al. Facilitators and barriers to quality of care in 46 maternal, newborn and child health: a global situational analysis through metareview. 47 Bmj Open 2014; 4(5): e004749. 48 19. Carroli G, Villar J, Piaggio G, et al. WHO systematic review of randomised controlled 49 trials of routine antenatal care. Lancet 2001; 357(9268): 1565-70. 50 20. Dowswell T, Carroli G, Duley L, et al. Alternative versus standard packages of 51 antenatal care for low-risk pregnancy. Cochrane Database Syst Rev 2015; (7): 52 CD000934. 53 54 21. Mortensen B. To be weiled or not to be - what unite is the question, Experiences from 55 a continuity of Midwifery Care Model in Palestine and Norway. Bodø, Norway: Nord 56 University; 2011. p. 121. 57 58 14 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 22. Villar.J BP. Antenatal Care Randomized Trial: Manual for the Implementation of the 4 new model. . http://apps.who.int/iris/bitstream/10665/42513/1/WHO_RHR_01.30.pdf 5 UNDP, UNFPA, WHO, World Bank Special Programme of Research Development 6 and Research in Human Reproduction, 2002. 7 23. Kearns AD, Caglia JM, ten Hoope-Bender P, Langer A. Antenatal and postnatal care: 8 a review of innovative models for improving availability, accessibility, acceptability 9 and quality of services in low-resource settings. Bjog-Int J Obstet Gy 2016; 123(4): 10 540-8. 11 24. WHO. Standards for improving quality of maternal and newborn care in health 12 facilities. http://apps.who.int/iris/bitstream/10665/249155/1/9789241511216- 13 14 eng.pdf?ua=1 WHO; 2016. 15 25. Tuncalp, Were WM, MacLennan C, et al. Quality of care for pregnant women and 16 newborns-theFor WHO peer vision. BJOG review 2015; 122(8): 1045-9. only 17 26. Oyibo PG, Ebeigbe PN, Nwonwu EU. Assessment of the risk status of pregnant 18 women presenting for antenatal care in a rural health facility in Ebonyi State, South 19 Eastern Nigeria. N Am J Med Sci 2011; 3(9): 424-7. 20 27. Hassan SJ, Wick L, DeJong J. A glance into the hidden burden of maternal morbidity 21 and patterns of management in a Palestinian governmental referral hospital. Women 22 Birth 2015; 28(4): e148-56. 23 28. WHO. Recommendations of postnatal care for mothers and new born, 2013 The 24 Departments of Maternal, Newborn, Child and Adolescent Health and Reproductive 25 Health and Research of the World Health Organization. 26 27 http://apps.who.int/iris/bitstream/10665/97603/1/9789241506649_eng.pdf?ua=1 28 World Health organization; 2013. p. 62. 29 29. Filby A, McConville F, Portela A. What Prevents Quality Midwifery Care? A 30 Systematic Mapping of Barriers in Low and Middle Income Countries from the 31 Provider Perspective. PLoS One 2016; 11(5): e0153391. 32 30. Peters DH, Adam T, Alonge O, Agyepong IA, Tran N. Implementation research: what http://bmjopen.bmj.com/ 33 it is and how to do it. BMJ 2013; 347: f6753. 34 31. Friedman HS, Liang M, Banks JL. Measuring the cost-effectiveness of midwife-led 35 versus physician-led intrapartum teams in developing countries. Womens Health 36 (Lond) 2015; 11(4): 553-64. 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 15 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 Clinics (clusters) in the 35 two regions (n=53) 36 total population: 528 054 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 Excluded: 43 Urban clinics (n=8) 44 Clinics45 with incomplete intervention period (n=3) 46 Clinics established after intervention (n=3) 47 48 49 50 51 52 Clinics included 53 54 in the study (n=39) 55 56 57 58 59 60 For peerClinics review with only - http://bmjopen.bmj.com/site/about/guidelines.xhtmlClinics in intervention (n=14) control group (n=25) Page 17Bizzariya of 20 BMJ Open Burqa Yasid Beit Imrin BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 Sabastiya Talluza 3 An Naqura Al Badhan 4 »º Deir Sharaf Asira ash Shamaliya 5 An Nassariya 6 Beit Iba 7 Qusin 8 Nablus Azmut 9 v® Deir al Hatab 10 Sarra 11 »º Tell 12 Rujeib Beit Dajan 13 Marj Na'ja Burin 14 Az Zubeidat 15 »º 16 ForAwarta peer review only 17 Urif 18 Einabus Huwwara 19 Beita 20 Jamma'in Al Jiftlik-abu al 'Ajaj 21 Aqraba 22 23 »º Yatma 24 Qabalan 25 »º 26 As Sawiya Qusra Majdal Bani Fadil 27 Salfit 28 Qaryut 29 30 Duma 31 32 http://bmjopen.bmj.com/ 33 34 Fasayil 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 !i Intervention Clinic 42 Control Clinic 43 Al 'Auja 44 v® Hospital 45 ¹º» Regular checkpoint 46 47 Palestinian Built-up Area 48 Israeli Military Base»º 49 50 IsraeliRamallah Settlement 51 Israeli Nature Reserve 52 53 Israeli Firing Zone 54 »º Oslo Agreement Areas 55»º »º Jericho »º 56 Area A »º 57 »º 58 Area B v® 59 Nature Reserve 60 »º For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml0 2.5 5 10 Oslo area C »º Kilometers BMJ Open Page 18 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 TREND Statement Checklist 2 3 Paper Item Descriptor Reported? 4 Section/ No 5 Pg # 6 Topic 7 Title and Abstract 8 9 Title and 1 Information on how unit were allocated to interventions 1 10 Abstract Structured abstract recommended 1 11 Information on target population or study sample 1 12 13 Introduction 14 Background 2 Scientific background and explanation of rationale 1 & 2 15 Theories used in designing behavioral interventions 16 For peer review only 2 17 Methods 18 19 Participants 3 Eligibility criteria for participants, including criteria at different levels in 3 & 4 20 recruitment/sampling plan (e.g., cities, clinics, subjects) 21 Method of recruitment (e.g., referral, self-selection), including the 22 sampling method if a systematic sampling plan was implemented 3 & 4 23 Recruitment setting 3 & 4 24 Settings and locations where the data were collected 3 & 4 25 Interventions 4 Details of the interventions intended for each study condition and how 26 3 27 and when they were actually administered, specifically including: 28 o Content: what was given? 3 29 o Delivery method: how was the content given? 3 30 o Unit of delivery: how were the subjects grouped during delivery? 31 o Deliverer: who delivered the intervention? 32 3 http://bmjopen.bmj.com/ 33 o Setting: where was the intervention delivered? 3 34 o Exposure quantity and duration: how many sessions or episodes or 35 events were intended to be delivered? How long were they 3 36 intended to last? 37 o Time span: how long was it intended to take to deliver the 38 intervention to each unit? 3 39 o Activities to increase compliance or adherence (e.g., incentives) 40 3 on September 29, 2021 by guest. Protected copyright. 41 Objectives 5 Specific objectives and hypotheses 3 42 Outcomes 6 Clearly defined primary and secondary outcome measures 4 43 Methods used to collect data and any methods used to enhance the 44 quality of measurements 4 45 Information on validated instruments such as psychometric and biometric 46 4 47 properties Sample Size 7 How sample size was determined and, when applicable, explanation of any 48 4 49 interim analyses and stopping rules 50 Assignment 8 Unit of assignment (the unit being assigned to study condition, e.g., 51 Method individual, group, community) 4 52 Method used to assign units to study conditions, including details of any 53 4 54 restriction (e.g., blocking, stratification, minimization) Inclusion of aspects employed to help minimize potential bias induced due 55 5 56 to non-randomization (e.g., matching) 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 20 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 TREND Statement Checklist 2 3 Blinding 9 Whether or not participants, those administering the interventions, and 4 (masking) those assessing the outcomes were blinded to study condition assignment; 5 if so, statement regarding how the blinding was accomplished and how it 6 was assessed. 7 8 9 Unit of Analysis 10 Description of the smallest unit that is being analyzed to assess 10 intervention effects (e.g., individual, group, or community) 11 5 12 If the unit of analysis differs from the unit of assignment, the analytical 13 method used to account for this (e.g., adjusting the standard error 5 14 estimates by the design effect or using multilevel analysis) 15 Statistical 11 Statistical methods used to compare study groups for primary methods 16 Methods Foroutcome(s), peer including complexreview methods of correlatedonly data 5 17 Statistical methods used for additional analyses, such as a subgroup 18 5 19 analyses and adjusted analysis 20 Methods for imputing missing data, if used 21 Statistical software or programs used 5 22 23 Results 24 Participant flow 12 Flow of participants through each stage of the study: enrollment, 25 assignment, allocation, and intervention exposure, follow-up, analysis (a 6 26 diagram is strongly recommended) 27 o 28 Enrollment: the numbers of participants screened for eligibility, 29 found to be eligible or not eligible, declined to be enrolled, and 6 30 enrolled in the study 31 o Assignment: the numbers of participants assigned to a study 6 32 condition http://bmjopen.bmj.com/ 33 o Allocation and intervention exposure: the number of participants 34 assigned to each study condition and the number of participants 6 35 who received each intervention 36 37 o Follow-up: the number of participants who completed the follow- 38 up or did not complete the follow-up (i.e., lost to follow-up), by 6 39 study condition 40 o Analysis: the number of participants included in or excluded from 6 on September 29, 2021 by guest. Protected copyright. 41 the main analysis, by study condition 42 Description of protocol deviations from study as planned, along with 43 reasons 44 45 Recruitment 13 Dates defining the periods of recruitment and follow-up 3 46 Baseline Data 14 Baseline demographic and clinical characteristics of participants in each 47 study condition 5,6 & 7 48 Baseline characteristics for each study condition relevant to specific 7 49 disease prevention research 50

51 Baseline comparisons of those lost to follow-up and those retained, overall 52 and by study condition Comparison between study population at baseline and target population 53 7 & 8 54 of interest 55 Baseline 15 Data on study group equivalence at baseline and statistical methods used 56 equivalence to control for baseline differences 8 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 20 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 TREND Statement Checklist 2 3 Numbers 16 Number of participants (denominator) included in each analysis for each 4 analyzed study condition, particularly when the denominators change for different 7 5 outcomes; statement of the results in absolute numbers when feasible 6 Indication of whether the analysis strategy was “intention to treat” or, if 5 7 not, description of how non-compliers were treated in the analyses 8 Outcomes and 17 For each primary and secondary outcome, a summary of results for each 9 estimation estimation study condition, and the estimated effect size and a confidence 10 8 11 interval to indicate the precision Inclusion of null and negative findings 12 8 13 Inclusion of results from testing pre-specified causal pathways through 14 15 which the intervention was intended to operate, if any Ancillary 18 Summary of other analyses performed, including subgroup or restricted 16 For peer review only 7,8&9 17 analyses analyses, indicating which are pre-specified or exploratory 18 Adverse events 19 Summary of all important adverse events or unintended effects in each 19 study condition (including summary measures, effect size estimates, and 8 20 confidence intervals) 21 22 DISCUSSION 23 Interpretation 20 24 Interpretation of the results, taking into account study hypotheses, 25 sources of potential bias, imprecision of measures, multiplicative analyses, 9,10, 26 and other limitations or weaknesses of the study 11&12 27 Discussion of results taking into account the mechanism by which the 28 intervention was intended to work (causal pathways) or alternative 9,10, 11&12 29 mechanisms or explanations 30 Discussion of the success of and barriers to implementing the intervention, 31 11 32 fidelity of implementation http://bmjopen.bmj.com/ 33 Discussion of research, programmatic, or policy implications 9,10,11&12 34 Generalizability 21 Generalizability (external validity) of the trial findings, taking into account 35 the study population, the characteristics of the intervention, length of 11 36 follow-up, incentives, compliance rates, specific sites/settings involved in 37 the study, and other contextual issues 38 Overall 22 39 General interpretation of the results in the context of current evidence 12 40 Evidence and current theory on September 29, 2021 by guest. Protected copyright. 41 42 43 From: Des Jarlais, D. C., Lyles, C., Crepaz, N., & the Trend Group (2004). Improving the reporting quality of 44 nonrandomized evaluations of behavioral and public health interventions: The TREND statement. American Journal of 45 46 Public Health, 94, 361-366. For more information, visit: http://www.cdc.gov/trendstatement/ 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

CAN A MIDWIFERY-LED CONTINUITY MODEL IMPROVE MATERNAL SERVICES IN A LOW RESOURCE SETTING? – A NON-RANDOMIZED CLUSTER INTERVENTION STUDY IN PALESTINE

ForJournal: peerBMJ Open review only

Manuscript ID bmjopen-2017-019568.R1

Article Type: Research

Date Submitted by the Author: 08-Dec-2017

Rikshospitalet, Oslo, Norway; Institute for clinical medicine, Faculty of http://bmjopen.bmj.com/ Medicine, University of Oslo, Oslo, Norway

Primary Subject Obstetrics and gynaecology Heading:

Secondary Subject Heading: Global health, Public health, Health services research, Mental health

Continuity of care, Midwifery, Organisation of health services < HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Quality in health care <

Keywords: on September 29, 2021 by guest. Protected copyright. HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Community child health < PAEDIATRICS, Maternal care

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 CAN A MIDWIFERY-LED CONTINUITY MODEL IMPROVE MATERNAL 4 SERVICES IN A LOW RESOURCE SETTING? – A NON-RANDOMIZED CLUSTER 5 6 INTERVENTION STUDY IN PALESTINE 7 1 ,2 3,4 5 2,6 8 Berit Mortensen MCs , Mirjam Lukasse PhD , Lien My Diep MSc , Marit Lieng PhD , 9 Amal Abu Awad, PhD7, Munjid Suleiman MCs8 Erik Fosse PhD1,2 10 11 12 Corresponding author: Berit Mortensen, Wilhelm Færdens vei 4b, 0361 Oslo, 13 14 [email protected] mobile: 0047-93266113 15 Abstract 16 For peer review only 17 Objectives To improve maternal health services in rural areas, the Palestinian Ministry of 18 Health launched a midwifery-led continuity model in the West Bank in 2013, in cooperation 19 20 with Norwegian Aid Committee. Midwives were weekly deployed from governmental 21 22 hospitals to provide ante- and postnatal-care in rural clinics. We studied the intervention`s 23 impact on utilization and quality indicators of maternal services after two years’ experience. 24 25 Design: A non-randomized, intervention design was chosen. The study was based on registry 26 27 data only available at cluster level, two years before (2011 & 2012) and two ears after (2014 28 & 2015) the intervention. 29 30 Setting: All 53 primary health care clinics in Nablus and Jericho regions were stratified for 31 inclusion. 32 http://bmjopen.bmj.com/ 33 Primary and secondary outcomes: Primary outcome was number of antenatal visits. 34 35 Important secondary outcomes were number of referrals to specialist care and number of 36 postnatal home-visits. Differences in changes within the two groups before and after the 37 38 intervention were compared by using mixed effect models. 39 Results: 14 intervention-clinics, and 25 control-clinics were included in the study. Number of 40 on September 29, 2021 by guest. Protected copyright. 41 antenatal visits increased by 1.16 per woman in the intervention-clinics while it declined by 42 43 0.39 in the control-clinics, giving a statistical significant difference in change of 1.55 visits 44 [95% CI 0.90 – 2.21]. A statistical significant difference in number of referrals was observed 45 46 between the groups giving a ratio of rate ratios of 3.65 [2.78 – 4.78] as number of referrals 47 48 increased by a rate ratio of 3.87 in the intervention group while in the control the rate-ratio 49 was only 1.06. 50 51 Home-visits increased substantially in the intervention group, but decreased in control group 52 giving a ratio of RR 97.65 (45.20 - 210.96) 53 54 55 56 57 58 1 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Conclusion: The Palestinian midwifery-led continuity model improved utilization and some 4 quality indicators of maternal services. More research should be done to investigate if the 5 6 model had effect on individual health outcomes and satisfaction of care. 7 8 Trial registration: ClinicalTrials.gov Identifier: NCT03145571 9 10 11 Strengths and Limitations of the study 12 • The pragmatic approach strengthens the applicability to real life settings. 13 14 • The high number of clusters and the robust cluster data strengthen the study. 15 16 • A randomizedFor allocation peer of clusters review was not possible onlybecause the implementation of 17 the midwifery-led continuity model started before the study was planned. 18 19 • The ministry implemented the program in the clinics they found appropriate which 20 21 could lead to bias. 22 • The facility-based registry did not include data on individual level. 23 24 INTRODUCTION 25 As a low-middle-income country under occupation, Palestine depends largely on foreign aid.1 26 27 The Palestinian Authority is responsible for the health services for the Palestinian people 28 29 within the occupied territories of West Bank and Gaza. In 2013, the Palestinian Ministry of 30 Health, registered 61 405 births and a fertility rate of 4.0 per woman on the West Bank.2 31 32 Maternal health services were provided by the Palestinian government and private, and non- http://bmjopen.bmj.com/ 33 34 governmental organisations. Less than 1% of the women gave birth at home. Governmental 35 facilities covered 45.6% of antenatal-care in 2013. The Palestinian Multiple Indicator Cluster 36 37 Survey from 2014 found that 66.4% of rural women gave birth in governmental hospitals. Of 38 them 40.7% left hospital within 6 hours postpartum and 73% did not receive any additional 39 40 3

postnatal-care. on September 29, 2021 by guest. Protected copyright. 41 42 In 2009 the Palestinian governmental maternal services were described as of poor quality, 43 overcrowded, understaffed, including short antenatal visits lacking content, and dissatisfaction 44 45 with care.4 Overcrowded labour rooms prohibited women bringing a birth companion.5 46 47 Midwives had restricted scope of practice and little autonomy, and few worked in antenatal- 48 care.6 The clinics were not able to implement postnatal home-visits as required by 49 6 7 50 governmental standards. Poor women were less likely to have postnatal-care. 51 Poverty, deprived infrastructure, military checkpoints and armed Israeli settlers restricts 52 53 freedom of movement and reduces access to central health-facilities and legal aid in rural 54 8,9 55 areas. During an escalation of the conflict between the years 2000 to 2006, it was reported 56 that 69 women gave birth at military checkpoints, causing casualties for both women and 57 58 2 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 4 3 children, as they were not allowed to pass to reach hospitals. Although the political situation 4 in the West Bank was less volatile the following decade, rural women are vulnerable and 5 6 depend more on governmental facilities than women in urban areas, as rural private services 7 10 8 are scarce. 9 Several studies describe how midwifery-led continuity models improved health for mothers 10 11-16 11 and babies. Most studies were from high-income countries. World Health Organisation 12 (WHO) recommends implementation and research on midwifery-led continuity models to 13 14 improve quality in low- and middle-income countries.17 Two main ways of organizing such 15 16 models are describedFor in the peer literature. In reviewthe case-load model only one midwife cares for up to 40 17 women and facilitate relational continuity, while in the team-midwifery model a group of 4-6 18 19 midwives can provide care for up to 360 women through pregnancy, birth and postnatally. 20 11,17 21 Ideally, in both models, women are cared for during birth by a known midwife. 22 To improve services in rural areas, the Palestinian Ministry of Health in cooperation with the 23 24 non-governmental humanitarian organization Norwegian Aid Committee (NORWAC), 25 launched implementation of a modified midwifery-led continuity, case-load-model, in 2013, 26 27 starting in Nablus and Jericho Governmental hospitals and surrounding villages. The 28 29 implementation engaged several levels in the ministry of health and the involved 30 communities, to overcome known barriers to quality of care.18 31 32 The aim of this study was to investigate whether the Palestinian midwifery-led continuity http://bmjopen.bmj.com/ 33 34 model had impact on the utilization of maternity services and selected quality indicators at the 35 two regions` clinics after two years of experience. 36 37 38 METHODS 39 40

Implementation of the Palestinian Midwifery-led continuity model on September 29, 2021 by guest. Protected copyright. 41 42 The modified case-load model aimed at establishing a relationship between the pregnant 43 woman and her midwife, during pregnancy and the postnatal period. The midwife also 44 45 worked at the governmental hospital, where most woman would give birth unless they chose a 46 47 private hospital. 48 Midwives in Nablus and Jericho governmental hospitals were assigned, trained, and took 49 50 driving lessons to obtain driving license. Under supervision, they provided ante-and postnatal 51 care in vicinity villages` governmental clinics. The hospital in Nablus had enough midwives 52 53 to serve ten villages per week, meaning two midwives would leave hospital each weekday, 54 55 five days a week. Jericho, served five villages, one midwife visited one village each day. The 56 hospitals were upscaled with three and two extra midwives respectively, to maintain the 57 58 3 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 capacity at the labour ward. All midwives worked full time, as part time employment was not 4 possible in the Ministry of Health. The same midwife visited the same village, usually once a 5 6 week. If the designated midwife had vacation or sick leave one of the other midwives would 7 8 cover her village. The ideal case load per midwife was around 50 pregnant women yearly, but 9 should not exceed 100. Thus, the smallest village with 16 registered pregnant women per year 10 11 was visited only every second week and the largest village with 163 registered pregnant 12 women was shared by two midwives weekly. The remaining working days the midwife spent 13 14 in the labour ward. All pregnant women were informed that their midwife during pregnancy 15 16 also worked at theFor local governmental peer hospital. review Independent onlyof birth place, all women 17 registered at the clinic should be offered postnatal home visits. All pregnant women were 18 19 informed that the limited numbers of midwives, and the large workload in the labour ward, 20 21 made it difficult to ensure a known midwife during labour. Women were given the phone 22 number for their midwife in case of an emergency. The level of relational continuity through 23 24 the case-load model in Palestine was overall limited to the ante- and postnatal period. 25 Nevertheless, a relational continuity was possible also during birth, if the known midwife was 26 27 on duty. Implementing the model aimed to strengthening the relationship between the woman 28 29 and her midwife, improve interdisciplinary cooperation, and reduce the barrier between 30 hospital and primary health care. The midwives received driving lessons, and used designated 31 32 cars with the Ministry of Health logo, and marked Midwifery Care, to facilitate transportation http://bmjopen.bmj.com/ 33 34 to villages and homes. Standard care in the clinics without this model was offered mainly 35 from nurses or midwives and medical doctors (GP) working only in primary health care. 36 37 38 Study design 39 40 As the implementation started before the study, a non-randomized intervention design was on September 29, 2021 by guest. Protected copyright. 41 42 chosen to evaluate the model, based on registry data at cluster level from two complete years 43 before (2011 & 2012), and two complete years after (2014 & 2015) the intervention. Clinics 44 45 where the model was implemented were compared with clinics where the model was not 46 47 implemented. Both arms of the study followed the same written governmental procedures. 48 The study was part of an implementation research project aiming at documenting the effect of 49 50 the midwifery-led continuity model. The study was approved by the Regional ethical 51 committee of South East Norway and by the Palestinian Ministry of Health. 52 53 54 55 56 57 58 4 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Clusters 4 The clusters consisted of governmental primary health village-clinics in Nablus and Jericho 5 6 regions. There were 53 active clinics during the study period. During autumn 2013, the 7 8 midwifery model had been implemented in 16 clinics. 9 All clinics were stratified for inclusion by rural and urban location, activity period, and 10 11 intervention period. 12

13 14 Exclusion criteria 15 16 Clinics located lessFor than three peer km from Nablusreview and Jericho centreonly were defined as urban, thus 17 pregnant women had better access to private and non-governmental services. Clinics in urban 18 19 areas were therefor excluded. Clinics opened during the study period were excluded due to 20 21 incomplete data. 22 Clinics where the intervention was prematurely terminated or introduced later, during the 23 24 study period were excluded because of contaminated data. 25 26 27 Outcomes 28 29 The number of antenatal visits was chosen as primary outcome. Secondary outcomes were 30 number of pregnant women referred to higher level of care, and number of women receiving 31 32 postnatal home-visits. Other outcomes were number of women registered at the clinic for http://bmjopen.bmj.com/ 33 34 antenatal-care, number of pregnant referred for abnormal blood sugar levels, number of 35 women seen by doctor after birth, number of newborn seen by doctor after birth and number 36 37 of total postnatal consultations for mother and newborn. 38 39 40

Statistical analysis on September 29, 2021 by guest. Protected copyright. 41 42 Aggregated data were retrieved from the governmental registry. The registry consisted of 43 anonymous recordings reported monthly from all clinics to the central statistical database in 44 45 the Ministry of Health. The registry did not include data at an individual level. 46 47 Mean, standard deviation and range were given for normally distributed and count variables. 48 Percentage and total number were given for categorical variables. 49 50 51 Change from baseline in the intervention and standard care groups, and any differences 52 53 between the groups changes, were examined by using mixed effects models. In the mixed 54 55 models, the clinic was specified as cluster (i.e. random variable), time and group, and 56 interaction between time and group were treated as fixed variables. 57 58 5 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 4 Approximately normally distributed count outcomes were fitted by mixed effects linear 5 6 regressions. Before fitting, the outcome variables were divided by the number of registered 7 8 pregnant women, or newborn, to calculate average values. Variance weights for each average 9 were then computed and included in a variance formula in the model because of 10 11 heteroscedasticity, since the computed averages are based on different numbers of pregnant 12 women. 13 14 15 16 Mixed effects PoissonFor regressions peer were usedreview to fit non-normally only distributed count outcomes, 17 and an offset variable was used to adjust for the total number of (individual) registrations that 18 19 were under risk in the models. 20 21 22 Measured confounding variables, which could have influenced the key estimates, were: the 23 24 villages population size, if the clinic had an employed community midwife and laboratory 25 equipment, regular military check point between village and hospital, and distant from 26 27 hospital. Those possible confounders were included in the mixed models for adjusting. 28 29 30 Adjusted regression coefficient as means and rate ratios with 95% confidence intervals were 31 32 given. Two-sided p-values of less than 0.05 was considered statistically significant. http://bmjopen.bmj.com/ 33 34 35 The map was developed by using ArcGIS software, and the attribute/database are part of the 36 37 same application. 38 39 40 The analyses of mixed effect models were performed with R version 3.4 and STATA version on September 29, 2021 by guest. Protected copyright. 41 42 14. Descriptive analyses were carried out using IBM-SPSS version 21 for Windows. 43 44 45 46 47 RESULTS 48 After stratification, 39 clinics were included in the study, 14 as intervention and 25 as control 49 50 clinics (figure 1). 51 In total, 10 034 women booked at the 39 included clinics during the study period, 2 784 in the 52 53 intervention clinics and 7 250 in the control clinics. 54 55 The clinics location is presented in figure 2. 56 57 58 6 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 The Palestinian Ministry of Health confirmed that no other activities were introduced 4 unequally to the groups during the study period. The clinics were in a region where political 5 6 unrest and economic hardship most likely would affect the intervention and control groups 7 8 similarly during the study period. The measured possible confounders presented in table 1 9 were adjusted for in the final results, none had significant confounding effect. 10 11 12 13 Table 1 Characteristics of clusters*

14 Characteristics Intervention Control 15 Population served mean 3402 4636 16 For peermin/max review 1000/7554 1875/11017 only 17 Distance to hospital (km) mean 23 12.6 18 min/max 5/59 3/28 19 20 Number of clinics with employed 0 8 21 community midwife 22 Number of clinics with laboratory 5 10 23 Additional clinics in village (NGO) 4 2 24 Number of clinic with regular military 6 14 25 checkpoints between village and hospital 26 * 14 clinics with intervention and 25 clinics with standard care(control) 27 28 29

30 31 Descriptive statistics of primary and secondary outcome variables such as number of 32 http://bmjopen.bmj.com/ 33 individuals registered, mean, standard deviation and range for the two timepoints in the 34 intervention and standard care are presented in table 2. 35 36 37 Table 2 Descriptive statistics for primary and secondary outcomes at baseline and two years after the 38 implementation 39 Groups/clusters Mean 40

(n=39)* timepoint** N*** **** SD Range on September 29, 2021 by guest. Protected copyright. 41 min max 42 Before 1094 39 18 12 95 43 Intervention 44 Number of women registered After 1690 60 33 16 163 for ANC during study period Before 3180 64 36 3 168 45 Standard care 46 After 4070 81 42 23 213 47 Total 10034 48 Before 2220 79 42 27 176 49 Intervention 50 Number of newborn registered After 2470 108 61 29 217 51 during study period Before 5416 88 46 33 291 Standard care 52 After 5771 115 67 36 298 53 total 15877 54 Number of antenatal visits and Before 4015 3.7 0.9 1.9 5.6 Intervention 55 mean visit per women per 56 After 7994 4.7 1.2 3.0 7.0 57 58 7 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 cluster (recurrent visits divided Before 14657 4.6 1.4 2.4 8.6 4 on number of registered Standard care women) 5 After 16769 4.2 1.1 2.4 6.8 6 Before 79 7.3 8.3 0 36.4 Number of referrals to higher Intervention 7 After 456 25.6 14.2 2.3 54.1 level of care (mean % of 8 Before 427 12.0 11.7 0 45.5 registered pregnant per cluster) Standard care 9 After 549 12.8 13.2 0 66.7 10 Before 12 1.7 4.7 0 17.7 11 Number of registered pregnant Intervention 12 women receiving home visits After 721 41.8 25.2 0 97.5 13 after birth (mean % of Before 42 1.5 4.9 0 25.5 Standard care 14 registered pregnant per cluster) After 22 0.7 2.3 0 11.5 15 Before 57 27 21 109 16 Coverage - ratio For% between ofpeer Intervention review only 17 number of registered newborns After 71 22 37 131 18 and registered pregnant per Before 61 23 7 119 19 cluster Standard care 20 After 74 16 47 113 Before 29 2.6 3.4 0 10.0 21 Number of referrals because of Intervention 22 abnormal blood sugar (mean % After 81 4.5 7.4 0 30.2 23 of registered pregnant per Before 90 2.6 3.5 0 14.9 cluster) Standard care 24 After 105 2.7 3.6 0 13.0 25 Before 208 12.3 20.0 0 0.7 26 Number of mothers seen by Intervention After 461 20.1 32.4 0 94.0 27 doctor postnatally (mean % of 28 Before 534 12.1 20.5 0 93.0 registered newborn per cluster) Standard care 29 After 225 4.8 10.0 0 57.0 30 Before 1670 79.6 24.5 29.7 118.6 31 Number of newborn seen by Intervention After 2173 91.4 29.8 34.8 172.4 32 doctor postnatally (mean % of Before 4338 85.2 26.8 21.8 162.2 http://bmjopen.bmj.com/ 33 registered newborn per cluster) Standard care 34 After 5082 90.4 21.6 47.5 142.1 35 Before 3902 1.8 0.3 1.3 2.2 Intervention 36 Total postnatal consultation for After 5364 2.2 0.4 1.7 3.1 37 newborn Before 9796 1.9 0.3 1.2 2.6 38 Standard care 39 After 10875 1.9 0.2 1.5 2.4 40 Before 1830 0.9 0.4 0.2 1.7

Intervention on September 29, 2021 by guest. Protected copyright. 41 Total postnatal consultations After 3637 1.4 0.4 0.8 2.1 42 for mothers of newborn Before 5073 1.0 0.5 0.2 3.1 Standard care 43 After 5399 1.0 0.7 0.2 1.6 44 45 * 14 intervention clusters and 25 Standard care clusters** Two years before intervention (2011 & 2012) and two years after intervention (2014 & 2015) *** Number of total individual registrations**** The mean at cluster level 46 47 48 Change within the intervention and standard care group as means and rate ratios (RRs), and 49 50 difference between the changes within the two groups, controlled for potential confounding 51 covariates, are presented in table 3. 52 53 54 55 56 57 58 8 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Table 3 Change before and after intervention in both groups and multiplicative difference of 4 changes between the groups Change in groups before and after Difference in 5 changes between 6 Outcome* Group groups P-value 7 Adjusted mean 8 Adjusted***mean (95%CI) (95% CI) 9 Mean number of Intervention 1.16 (0.60 - 1.72) p<0.0001 10 antenatal visits per 1.55 (0.90 - 2.21) 0.0007

11 pregnant Control -0.39 (-0.73 - 0.05) p=0.026 12 13 Number of newborn Intervention 18.2% (10.0 - 26.4) p<0.0001 mothers who 14 6.6% (-3.1 - 16.4) 0.179 registered at clinic Control 15 during their pregnancy 11.6% (6.3 - 16.9) p<0.0001 16 For peer review only 17 Number of newborn Intervention 18 12.8% (-1.8 - 27.3) p=0.085 seen by doctor 4.7% (-12.7 - 21.9) 0.599 19 postnatally Control 8.1% (-1.3 - 17.5) p=0.089 20

21 Total number of Intervention 22 0.41 (0.26 - 0.57) p<0.0001 postnatal consultations 0.33 (0.16 - 0.52) 0.0004 23 for newborn Control 0.08 (-0.02 -0.18) p=0.126 24 25 Total number postnatal Intervention 0.64 (0.52 - 0.77) p<0.0001 26 consultations for 0.60 (0.46 - 0.75) <0.0001 27 mothers Control 0.04 (-0.04 -0.12) p=0.321 28 29 Adjusted RR (95% Outcome** Group P-value 30 Adjusted Rate Ratio RR (95%CI) CI) 31 32 Intervention 3.87 (3.04 - 4.92) p <0.0001 Number of referrals to http://bmjopen.bmj.com/ 3.64 (2.78 - 4.78) <0.0001 33 higher level of care Control 34 1.06 (0.94 - 1.21) p=0.353 35 Number of registered 36 Intervention 37.42 (21.14 - 66.22) p<0.0001 97.65 (45.20 - pregnant who received <0.0001 37 210.96) postnatal home visits Control 0.38 (0.23 - 0.64) p<0.0001 38

39 Number of referrals 40 Intervention 1.78 (1.16 - 2.72) p= 0.008 because of abnormal 1.83 (1.10 - 3.05) 0.021 on September 29, 2021 by guest. Protected copyright. 41 blood sugar Control 0.97 ((0.73-1.29) p=0.846 42 43 Number of mothers Intervention 1.94 (1.65 - 2.29) p<0.0001 44 seen by doctor 4.87 (3.88 - 6.10) <0.0001 45 postnatally Control 0.40 (0.34 - 0.47) p<0.0001 46 47 *Mixed effect linear regression was used to analyse change in and between clusters when data had normal 48 distribution**Mixed-effect Poisson regression was used to analyse the change in clusters when data was not normal 49 distributed8 ***All outputs were adjusted for potential confounders without any change in value. Covariates 50 adjusted for were: Distance from clinic to city hospital (km), population in village, if there were additional clinic in 51 village, clinics with community midwife, clinics with laboratory and clinics with a regular checkpoint on the way to 52 hospital. 53 54 55 56 Antenatal utilization 57 58 9 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 There was statistically significant difference in average change in mean number of antenatal 4 visits between the groups by 1.55 [1.38 – 1.54], p=0.0004, mean number of visits increased 5 6 by 1.16 visits with the new model, while standard care declined with - 0.39 visits. In other 7 8 words, clinics with the new model had an increase from 3.7 mean number of antenatal visits 9 per pregnant woman before the model was introduced to 4.7 mean number of antenatal visits 10 11 per woman after. While in the control clinics, mean number of antenatal visits per woman 12 decreased from 4.6 to 4.2 visits. 13 14 15 16 For peer review only 17 Referrals 18 19 A statistically significant difference in change between the groups` number of referrals to a 20 21 higher level of care was observed giving a ratio of rate-ratios of 3.64 [2.78 - 4.78], p<0.0001. 22 For the intervention group referrals increased by a rate-ratio of 3.87, meaning that the number 23 24 of referrals increased from 7.3% to 25.6% of all registered women in the clinics, while the 25 control group only had a change rate-ratio of 1.06, meaning that the percentage of referrals 26 27 moved only from 12 to 12.8% during the study period. 28 29 30 Postnatal service 31 32 Postnatal home-visits increased substantially at the intervention-clinics, while at the control- http://bmjopen.bmj.com/ 33 34 clinics it dropped giving a ratio of rate-ratios of 97.65 [45.20 – 210.96], p<0.0001. 35 Women in the intervention group increased significantly mean number of postnatal contacts 36 37 with health services, while no increase was observed at clinics with standard care, giving a 38 ratio of rate -ratios of 0.60 (95% CI 0.46 - 0.75) P<0.0001. Also, a significant increase in 39 40 mean number of newborn`s healthcare-contacts were observed with the new model, while not on September 29, 2021 by guest. Protected copyright. 41 42 in group with standard care giving a ratio of 0.33 (0.16 - 0.52) p=0.0004. 43 44 45 46 47 DISCUSSION 48 In the clinics with the midwifery-led continuity intervention a significant rise in mean number 49 50 of antenatal visits per woman was observed, whereas number of visits per woman decreased 51 in the clinics with standard care during the same period. It is thus likely that the improved 52 53 utilization was a result of the intervention. WHO recommended in 2002 a minimum of four 54 19 55 focused antenatal visits for healthy pregnant women. After evaluating new evidence, the 56 recommendation was revised in 2016 to a minimum of eight antenatal visits to reduce 57 58 10 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 17,18,20 3 perinatal mortality and improve women’s satisfaction. The women’s increased 4 adherence to service in clinics with midwifery-led continuity may indicate that women 5 6 experienced improved quality of the services. Relational continuity is an important tool to 7 16 8 enhance communication and thus satisfaction of care. The association between improved 9 quality of care and increased utilization is supported by several previous studies, and by WHO 10 11 recommendations, to improve utilization and quality by introducing midwife-led continuity of 12 care.12,14,17,18 13 14 It is a possibility that the pregnant woman would feel safe knowing that the midwife 15 16 following her throughoutFor pregnancypeer also review worked at the public only hospital where she would give 17 birth, and that her midwife would visit her at home after birth. Due to the heavy workload, the 18 19 midwives could not always be on call to attend birth. Midwifery-led continuity of care in 20 21 settings with few midwife resources and heavy workload, must balance the demands on the 22 midwives. Although women were not guaranteed that the same midwife providing them ante- 23 24 and postnatal care would attend their birth, their midwife`s connection to the governmental 25 hospital might have reduced the alien barrier to the hospital, and restored feeling of security 26 27 for the rural women. 28 29 A qualitative study investigated midwives’ experience from working in a similar model in 30 Ramallah region from 2007 to 2011. The midwives described how the model enabled them to 31 32 give individualized care and how the broader scope of practice and increased autonomy gave http://bmjopen.bmj.com/ 33 21 34 them important experience and tools to help. This can explain women adherence to service, 35 because building relationship with competent, respectful and motivated midwives probably 36 37 increased their wish to return to receive more care. 38 39 40 In the intervention group a change in referral mechanisms was observed. The midwives on September 29, 2021 by guest. Protected copyright. 41 42 working in the new model identified significantly more risk factors causing referral to higher 43 level of care, than in clinics with standard care. The identification of more women with 44 45 abnormal blood sugar level indicates that the model led to identification of important risk 46 47 factors. The proportion of pregnant women referred to higher level of care increased to 25.6% 48 with the new model vs. 13% with standard care during the study period. The proportion in the 49 50 new model is in line with the WHO-estimate presented in guidelines from 2001, that in 51 general 25% of pregnant women would need additional antenatal-care due to health 52 53 complications before or during pregnancy.22 The village clinics had little, if any, technical 54 55 resources to investigate risk signs, so referral to higher level of care was necessary to follow 56 up any possible complications. 57 58 11 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 One important quality indicator of antenatal-care is the ability to detect possible 4 complications and involve specialist care when necessary. Kearns et al previously 5 6 demonstrated that improved referral networks was a key element for improving quality in 7 23 8 low-resource settings. The process of information and referral within the system is also 9 highlighted by the WHO framework as one core indicator of quality of services.24,25 Some 10 11 countries have much higher potential health risks than others due to poverty, high fertility and 12 general health challenges.10,26 A Palestinian study from 2015 revealed that 26.9% of women 13 14 who gave birth experienced one or more morbidities.27 The rise in numbers of referrals after 15 16 introducing the newFor model peer matched the WHOreview and the local estimates.only Thus, it seems 17 reasonable to suggest that the intervention improved the referral system. 18 19 20 21 The results furthermore showed a substantial increase in number of postnatal-care contacts for 22 mothers and new-borns, including home-visits. WHO recommendation for postnatal care is a 23 24 minimum of three postnatal contacts; and minimum one home-visit preferably during the first 25 week after birth.28 The result from the study consequently indicates that the implementation of 26 27 midwifery-led continuity models may contribute to reach such goal. 28 29 Filby et al described how lack of transportation hampered quality improvements in other rural 30 resource-constrained settings.29 The implementation in rural Palestine included a designated 31 32 vehicle and driving skills for the midwives, facilitation of transportation was consequently a http://bmjopen.bmj.com/ 33 34 key in order to reach out to the villages and home-visits. 35 The fact that women receiving midwifery-led continuity were more frequently seen by a 36 37 doctor in the clinic after birth (20.1% vs. 4.8%), in addition to the midwife, indicate improved 38 interdisciplinary cooperation. When midwives undertook home-visit and discovered health 39 40 problems or risks, she involved the medical doctor. The findings also showed that there is a on September 29, 2021 by guest. Protected copyright. 41 42 systematic check of newborn by doctors in all the village clinics, and the home visit from the 43 midwife ads to this. The increase in and variation of, postnatal contacts including home-visits, 44 45 make it reasonable to conclude that the midwifery model improved both utilization and 46 47 quality of postnatal care at a cluster level. 48 49 50 Limitations of the study 51 The study was carried out after the implementation of the midwifery-led continuity model 52 53 started. This prevented a randomized allocation of clinics to intervention and control clusters. 54 55 The number of midwives available in the hospitals limited the number of clinics for 56 implementation to a total of 15 in these regions. The ministry chose to implement the program 57 58 12 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 in the clinics they found appropriate. The baseline data indicate that the reason for choosing 4 these clinics were challenges in service provision, thus improvements could be easier 5 6 achieved and lead to bias. 7 8 Another limitation was weak data due to the facility-based registry did not include data on 9 individual level. This is a common problem in low-and middle income-countries.23 Lack of 10 11 individual reproductive health registry prevented measuring individual impact and an intra- 12 cluster coefficient. It was not possible to identify when women registered at the clinic, who 13 14 came back for recurrent antenatal-visits or reasons for referrals. 15 16 For peer review only 17 Strength of the study and further recommendations 18 19 The high number of clusters and the robust cluster data strengthen the study. The 20 21 organizational leadership, engagement and adherence of the multidisciplinary team strengthen 22 the sustainably of a complex intervention and its applicability to real life settings. The 23 24 findings make it reasonable to conclude that the new model had effect on the utilization and 25 some quality indicators of the maternal service. The study can be a useful tool in power- 26 27 calculations and planning of randomized trials for future implementation of the model. 28 29 30 Triangulation of methods within an implementation research framework would be useful to 31 32 investigate the broader effect of the implementation. This is highly recommended when http://bmjopen.bmj.com/ 33 34 introducing evidence-based interventions to improve health service delivery in real world 35 settings where context is an important factor.30 Further research should be done to investigate 36 37 if the model has impact on individual health outcomes for mother and newborn, and 38 satisfaction of care. Previous research described midwifery-led models as cost saving, also in 39 40 developing countries.12,31 A study of this model`s cost-effectiveness would be useful. The on September 29, 2021 by guest. Protected copyright. 41 42 general understaffing in both primary and secondary governmental health services should be 43 taken in consideration. This demanded upscaling number of midwives to improve quality, at 44 45 some cost. The extra midwives were employed at the hospitals to serve the community, at the 46 47 same time the hospital got more midwives to share the shifts. Also, the midwives widened 48 their scope of practice and experience, while linking the primary and secondary healthcare. 49 50 The cost of transportation was reduced to a minimum by enabling the midwives to drive the 51 vehicle themselves, employing drivers would have added unsustainable cost to the model. 52 53 54 55 Conclusion 56 57 58 13 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 The findings make it reasonable to conclude that the new model had effect on the utilization 4 and some quality indicators of the maternal service. 5 6 The positive change in facility-level outcomes show that clinics with the midwifery model in 7 8 the regions of Nablus and Jericho improved services during pregnancy and postnatally. The 9 findings indicate improvement of utilization and some quality indicators linked to facility- 10 11 level outcomes, such as continuity, functioning referral system and postnatal home-visits. 12 The results from this study support the expansion of the model to new areas in Palestine. We 13 14 believe the model can be useful for other low- and middle-income countries to improve 15 16 utilization and qualityFor of care. peer review only 17 18 19 Authors affiliation 20 21 1 The Intervention Centre, Oslo University Hospital, Oslo, Norway, 2 Institute for Clinical 22 3 23 Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway, Faculty of Health 24 Sciences, Oslo and Akershus University College, Oslo, Norway, 4 Faculty of Health and 25 26 Social Sciences, University College of Southeast Norway, Oslo, Norway 5 Oslo Centre for 27 6 28 Biostatistics and Epidemiology, Oslo University Hospital, Oslo, Norway Department of 29 7 Gynaecology, Oslo University Hospital, Oslo, Norway, Palestinian Ministry of Health, 30 8 31 Nablus. Palestine Palestinian Ministry of Education and Higher Education, Ramallah, 32 Palestine http://bmjopen.bmj.com/ 33 34 35 36 Acknowledgements The authors would like to thank the Palestinian Ministry of Health for 37 their efforts in implementing the model, and preparing and sharing of their health registry. We 38 39 also like to thank Norwegian Aid Committee for supporting the study, and United Nations, 40 on September 29, 2021 by guest. Protected copyright. 41 Office for the coordination of Humanitarian Affairs (OCHA) for developing the map. Finally, 42 we are grateful to all the involved Palestinian midwives for their courageous efforts in 43 44 improving care for women in Palestine. 45

46 47 Contributors 48 49 BM was involved with the Implementation, study design, data collection, data analysis, data 50 interpretation and writing, MiL with study design, data interpretation and writing, LMD with 51 52 study design, data analysis and writing, MaL was involved with study design, data 53 54 interpretation and writing, AAA with the implementation, data interpretation and writing 55 56 57 58 14 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 MS conducted collection and systematization of data, EF involved in study design, data 4 collection, data analysis, data interpretation and writing. BM drafted the article, figures and 5 6 tables all authors have reviewed and approved the final manuscript. 7 8 Funding 9 This work was partly supported by the Research Council of Norway through the Global 10 11 Health and Vaccination Program (GLOBVAC), project number 243706 and partly by public 12 funding through Norwegian Aid Committee. Three of the Authors were partly employed by 13 14 NORWAC and were involved in the implementation and interpretation of data. The analysis 15 16 was performed byFor a statistician peer at Oslo Universityreview hospital. Theonly corresponding author had full 17 access to all the data in the study and had final responsibility for the decision to submit for 18 19 publication. 20 21 22 Competing interests None declared 23 24 25 Ethics approval The study was approved by the Norwegian Regional Committee for Medical 26 27 Health research Ethics South East (REK) id number: 2015/1235. It was also approved by the 28 29 Palestinian Ministry of Health. 30 Data sharing statement Data was obtained from the Palestinian Ministry of Health registry. 31 32 Figure legends: http://bmjopen.bmj.com/ 33 34 Figure 1. Flowchart showing the stratification of all clusters in the study area. 35 36 Figure 2. Map showing the location of all included clusters in the study area. 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 1. World Bank website Country information, accessed September 2017 43 http://data.worldbank.org/country/west-bank-and-gaza 44 2. Palestine Ministry of Health Annual report 2013. Palestine,Nablus, Palestine Ministry 45 of Health, 2013. 46 3. Statistics PCBo. Palestinian Multiple Indicator Cluster Survey 2014. Ramallah, 47 Palestine UNFPA, UNICEF, 2015. 48 4. Rahim HF, Wick L, Halileh S, et al. Maternal and child health in the occupied 49 Palestinian territory. Lancet 2009; 373(9667): 967-77. 50 5. Wick L, Mikki N, Giacaman R, Abdul-Rahim H. Childbirth in Palestine. Int J 51 Gynaecol Obstet 2005; 89(2): 174-8. 52 6. Abu Awad, Amal. PhD dissertation, Implementation of postnatal protocols by nurse in 53 54 primary health care clinics in the West Bank: University of Wisconsin-Madison 2011. 55 7. Nabaa HAA, Hilal GA, Sbeih SA, Ghandour R, Giacaman R. Access to care for 56 women reporting postnatal complications in the occupied Palestinian territory: a cross- 57 58 15 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 22

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 sectional study. The Lancet 2013; 382: S2. 4 8. Reem Al-Botmeh U. A review of Palestinian legislation from a women`s right 5 perspective. UNDP-papp-research-Legislative english.pdf UNDP, 2012. 6 9. United Nations, Office for the Coordination of Humanitarian Affairs occupied 7 Palestinian territory Area C Vulnerability Profile. http://data.ochaopt.org/vpp.aspx 8 Accessed September 2017. 9 10. Graham W, Woodd S, Byass P, et al. Diversity and divergence: the dynamic burden of 10 poor maternal health. Lancet 2016; 388(10056): 2164-75. 11 11. Homer C, Brodie P, Leap N. Midwifery continuity of care : a practical guide. Sydney; 12 New York: Churchill Livingstone/Elsevier; 2008. 13 14 12. Sandall J, Soltani H, Gates S, Shennan A, Devane D. Midwife-led continuity models 15 versus other models of care for childbearing women. Cochrane Database Syst Rev 16 2016; 4: ForCD004667. peer review only 17 13. ten Hoope-Bender P, de Bernis L, Campbell J, et al. Improvement of maternal and 18 newborn health through midwifery. Lancet 2014; 384(9949): 1226-35. 19 14. Renfrew MJ, McFadden A, Bastos MH, et al. Midwifery and quality care: findings 20 from a new evidence-informed framework for maternal and newborn care. Lancet 21 2014; 384(9948): 1129-45. 22 15. Homer CS, Friberg IK, Dias MA, et al. The projected effect of scaling up midwifery. 23 Lancet 2014; 384(9948): 1146-57. 24 16. Homer CS. Challenging midwifery care, challenging midwives and challenging the 25 system. Women Birth 2006; 19(3): 79-83. 26 27 17. WHO. WHO recommendations on antenatal care for a positive pregnancy experience. 28 http://apps.who.int/iris/bitstream/10665/250796/1/9789241549912-eng.pdf?ua=1 29 World Health Organization; 2016. p. 152. 30 18. Nair M, Yoshida S, Lambrechts T, et al. Facilitators and barriers to quality of care in 31 maternal, newborn and child health: a global situational analysis through metareview. 32 Bmj Open 2014; 4(5): e004749. http://bmjopen.bmj.com/ 33 19. Carroli G, Villar J, Piaggio G, et al. WHO systematic review of randomised controlled 34 trials of routine antenatal care. Lancet 2001; 357(9268): 1565-70. 35 20. Dowswell T, Carroli G, Duley L, et al. Alternative versus standard packages of 36 antenatal care for low-risk pregnancy. Cochrane Database Syst Rev 2015; (7): 37 CD000934. 38 21. Mortensen B. To be weiled or not to be - what unite is the question, Experiences from 39 a continuity of Midwifery Care Model in Palestine and Norway. Bodø, Norway: Nord 40 on September 29, 2021 by guest. Protected copyright. 41 University; 2011. p. 121. 42 22. Villar.J BP. Antenatal Care Randomized Trial: Manual for the Implementation of the 43 new model. . http://apps.who.int/iris/bitstream/10665/42513/1/WHO_RHR_01.30.pdf 44 UNDP, UNFPA, WHO, World Bank Special Programme of Research Development 45 and Research in Human Reproduction, 2002. 46 23. Kearns AD, Caglia JM, ten Hoope-Bender P, Langer A. Antenatal and postnatal care: 47 a review of innovative models for improving availability, accessibility, acceptability 48 and quality of services in low-resource settings. Bjog-Int J Obstet Gy 2016; 123(4): 49 540-8. 50 24. WHO. Standards for improving quality of maternal and newborn care in health 51 facilities. http://apps.who.int/iris/bitstream/10665/249155/1/9789241511216- 52 eng.pdf?ua=1 WHO; 2016. 53 54 25. Tuncalp, Were WM, MacLennan C, et al. Quality of care for pregnant women and 55 newborns-the WHO vision. BJOG 2015; 122(8): 1045-9. 56 26. Oyibo PG, Ebeigbe PN, Nwonwu EU. Assessment of the risk status of pregnant 57 58 16 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 22 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 women presenting for antenatal care in a rural health facility in Ebonyi State, South 4 Eastern Nigeria. N Am J Med Sci 2011; 3(9): 424-7. 5 27. Hassan SJ, Wick L, DeJong J. A glance into the hidden burden of maternal morbidity 6 and patterns of management in a Palestinian governmental referral hospital. Women 7 Birth 2015; 28(4): e148-56. 8 28. WHO. Recommendations of postnatal care for mothers and new born, 2013 The 9 Departments of Maternal, Newborn, Child and Adolescent Health and Reproductive 10 Health and Research of the World Health Organization. 11 http://apps.who.int/iris/bitstream/10665/97603/1/9789241506649_eng.pdf?ua=1 12 World Health organization; 2013. p. 62. 13 14 29. Filby A, McConville F, Portela A. What Prevents Quality Midwifery Care? A 15 Systematic Mapping of Barriers in Low and Middle Income Countries from the 16 Provider ForPerspective. peer PLoS One 2016;review 11(5): e0153391. only 17 30. Peters DH, Adam T, Alonge O, Agyepong IA, Tran N. Implementation research: what 18 it is and how to do it. BMJ 2013; 347: f6753. 19 31. Friedman HS, Liang M, Banks JL. Measuring the cost-effectiveness of midwife-led 20 versus physician-led intrapartum teams in developing countries. Womens Health 21 (Lond) 2015; 11(4): 553-64. 22 23 24 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 17 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 22 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Figure 1. Flowchart showing the stratification of all clusters in the study area. http://bmjopen.bmj.com/ 33 34 174x136mm (300 x 300 DPI) 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 Figure 2. Map showing the location of all included clusters in the study area. 46 47 209x297mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 22 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 TREND Statement Checklist 2 3 Paper Item Descriptor Reported? 4 Section/ No 5 Pg # 6 Topic 7 Title and Abstract 8 9 Title and 1 Information on how unit were allocated to interventions 1 10 Abstract Structured abstract recommended 1 11 Information on target population or study sample 1 12 13 Introduction 14 Background 2 Scientific background and explanation of rationale 1 & 2 15 Theories used in designing behavioral interventions 16 For peer review only 2 17 Methods 18 19 Participants 3 Eligibility criteria for participants, including criteria at different levels in 3 & 4 20 recruitment/sampling plan (e.g., cities, clinics, subjects) 21 Method of recruitment (e.g., referral, self-selection), including the 22 sampling method if a systematic sampling plan was implemented 3 & 4 23 Recruitment setting 3 & 4 24 Settings and locations where the data were collected 3 & 4 25 Interventions 4 Details of the interventions intended for each study condition and how 26 3 27 and when they were actually administered, specifically including: 28 o Content: what was given? 3 29 o Delivery method: how was the content given? 3 30 o Unit of delivery: how were the subjects grouped during delivery? 31 o Deliverer: who delivered the intervention? 32 3 http://bmjopen.bmj.com/ 33 o Setting: where was the intervention delivered? 3 34 o Exposure quantity and duration: how many sessions or episodes or 35 events were intended to be delivered? How long were they 3 36 intended to last? 37 o Time span: how long was it intended to take to deliver the 38 intervention to each unit? 3 39 o Activities to increase compliance or adherence (e.g., incentives) 40 3 on September 29, 2021 by guest. Protected copyright. 41 Objectives 5 Specific objectives and hypotheses 3 42 Outcomes 6 Clearly defined primary and secondary outcome measures 4 43 Methods used to collect data and any methods used to enhance the 44 quality of measurements 4 45 Information on validated instruments such as psychometric and biometric 46 4 47 properties Sample Size 7 How sample size was determined and, when applicable, explanation of any 48 4 49 interim analyses and stopping rules 50 Assignment 8 Unit of assignment (the unit being assigned to study condition, e.g., 51 Method individual, group, community) 4 52 Method used to assign units to study conditions, including details of any 53 4 54 restriction (e.g., blocking, stratification, minimization) Inclusion of aspects employed to help minimize potential bias induced due 55 5 56 to non-randomization (e.g., matching) 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

51 Baseline comparisons of those lost to follow-up and those retained, overall 52 and by study condition Comparison between study population at baseline and target population 53 7 & 8 54 of interest 55 Baseline 15 Data on study group equivalence at baseline and statistical methods used 56 equivalence to control for baseline differences 8 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 22 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

CAN A MIDWIFE-LED CONTINUITY MODEL IMPROVE MATERNAL SERVICES IN A LOW RESOURCE SETTING? – A NON-RANDOMIZED CLUSTER INTERVENTION STUDY IN PALESTINE.

ForJournal: peerBMJ Open review only

Manuscript ID bmjopen-2017-019568.R2

Article Type: Research

Date Submitted by the Author: 25-Jan-2018

Complete List of Authors: Mortensen, Berit; Oslo University Hospital, Rikshospitalet, The Intervention Center; University of Oslo, Faculty of Medicine, Institute for clinical Medicine Lukasse, Mirjam; Oslo Metropolian University, Faculty of Health Sciences, Department of Nursing and Health Promotion; University College of Southeast Norway, Department of Health and Social Science Diep, Lien; University of Oslo, International Health; Oslo University Hospital, Unit for biostatistics and epidemiology Lieng, Marit; Oslo University Hospital, Ullevål, Department of Obstetrics Awad, Amal; Palestinian Ministry of Health Suleiman, Munjid; Palestinian Ministry of Education and Higher Education

Fosse, Erik; The Intervention Centre, Oslo University Hospital http://bmjopen.bmj.com/ Rikshospitalet, Oslo, Norway; Institute for clinical medicine, Faculty of Medicine, University of Oslo, Oslo, Norway

Primary Subject Obstetrics and gynaecology Heading:

Secondary Subject Heading: Global health, Public health, Health services research, Mental health

Continuity of care, Midwifery, Organisation of health services < HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Quality in health care < on September 29, 2021 by guest. Protected copyright. Keywords: HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Maternal care, Low- Middle Income Country

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 CAN A MIDWIFE-LED CONTINUITY MODEL IMPROVE MATERNAL SERVICES 4 IN A LOW RESOURCE SETTING? – A NON-RANDOMIZED CLUSTER 5 6 INTERVENTION STUDY IN PALESTINE 7 1 ,2 3,4 5 2,6 8 Berit Mortensen MCs , Mirjam Lukasse PhD , Lien My Diep MSc , Marit Lieng PhD , 9 Amal Abu Awad, PhD7, Munjid Suleiman MCs8 Erik Fosse PhD1,2 10 11 12 Corresponding author: Berit Mortensen, Wilhelm Færdens vei 4b, 0361 Oslo, 13 14 [email protected] mobile: 0047-93266113 15 Authors affiliation 16 For peer review only 17 1 The Intervention Centre, Oslo University Hospital, Oslo, Norway, 2 Institute for Clinical 18 19 Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway, 3 Faculty of Health 20 4 21 Sciences, Oslo Metropolitan University, Oslo, Norway, Faculty of Health and Social 22 Sciences, University College of Southeast Norway, Oslo, Norway 5 Oslo Centre for 23 6 24 Biostatistics and Epidemiology, Oslo University Hospital, Oslo, Norway Department of 25 Gynaecology, Oslo University Hospital, Oslo, Norway, 7 Palestinian Ministry of Health, 26 27 Nablus. Palestine 8 Palestinian Ministry of Education and Higher Education, Ramallah, 28 29 Palestine 30 31 32 Abstract http://bmjopen.bmj.com/ 33 34 Objectives: To improve maternal health services in rural areas, the Palestinian Ministry of 35 Health launched a midwife-led continuity model in the West Bank in 2013. Midwives were 36 37 deployed weekly from governmental hospitals to provide antenatal and postnatal care in rural 38 clinics. We studied the intervention’s impact on utilization and quality indicators of maternal 39 40 services after two years’ experience. on September 29, 2021 by guest. Protected copyright. 41 42 Design: A non-randomized intervention design was chosen. The study was based on registry 43 data only available at cluster level, two years before (2011 & 2012) and two years after (2014 44 45 & 2015) the intervention. 46 47 Setting: All 53 primary health care clinics in Nablus and Jericho regions were stratified for 48 inclusion. 49 50 Primary and secondary outcomes: Primary outcome was number of antenatal visits. 51 Important secondary outcomes were number of referrals to specialist care and number of 52 53 postnatal home visits. Differences in changes within the two groups before and after the 54 55 intervention were compared by using mixed effect models. 56 57 58 1 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Results: 14 intervention-clinics, and 25 control-clinics were included. Number of antenatal 4 visits increased by 1.16 per woman in the intervention-clinics, while declined by 0.39 in the 5 6 control-clinics, giving a statistically significant difference in change of 1.55 visits [95% CI 7 8 0.90 – 2.21]. A statistically significant difference in number of referrals was observed 9 between the groups, giving a ratio of rate ratios of 3.65 [2.78 – 4.78] as number of referrals 10 11 increased by a rate ratio of 3.87 in the intervention group, while in the control the rate ratio 12 was only 1.06. 13 14 Home visits increased substantially in the intervention group, but decreased in the control 15 16 group, giving a ratioFor of RR peer 97.65 (45.20 review- 210.96) only 17 Conclusion: The Palestinian midwife-led continuity model improved utilization and some 18 19 quality indicators of maternal services. More research should be done to investigate if the 20 21 model influenced individual health outcomes, and satisfaction with care. 22 Trial registration: ClinicalTrials.gov Identifier: NCT03145571 23 24 25 Keywords: Midwifery, Continuity of care, Maternal Care, Quality in Health Care, 26 27 Organization of Health services, Less-developed country 28 Strengths and Limitations of the study 29 30 • The pragmatic approach strengthens the applicability to real life settings. 31 32 • The high number of clusters and the robust cluster data strengthen the study. http://bmjopen.bmj.com/ 33 • A randomized allocation of clusters was not possible because the implementation of 34 35 the midwife-led continuity model started before the study was planned. 36 37 • The ministry implemented the program in the clinics they found appropriate which 38 could have led to bias. 39 40 • The facility-based registry did not include data at individual level. on September 29, 2021 by guest. Protected copyright. 41 42 INTRODUCTION 43 As a low-middle-income country under occupation, Palestine depends largely on foreign aid.1 44 45 The Palestinian Authority is responsible for Palestinian health services in the occupied 46 47 territories of the West Bank and Gaza. In 2013, the Palestinian Ministry of Health registered 48 61 405 births and a fertility rate of 4.0 per woman in the West Bank.2 Maternal health services 49 50 were provided by the Palestinian government, and by private, and non-governmental 51 organisations. Less than 1% of the women give birth at home. Governmental facilities 52 53 covered 45.6% of antenatal care in 2013. The Palestinian Multiple Indicator Cluster Survey 54 55 from 2014 found that 66.4% of rural women gave birth in governmental hospitals. Of these 56 57 58 2 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 40.7% left hospital within 6 hours postpartum and 73% did not receive any additional 4 postnatal care.3 5 6 In 2009 the Palestinian governmental maternal services were described as of poor quality due 7 8 to concerns for being overcrowded and understaffed. Patient reported dissatisfaction with care 9 as antenatal visits were short and lacking content.4 Overcrowded labour rooms prohibited 10 5 11 women from bringing a birth companion. Midwives had restricted scope of practice and little 12 autonomy, and were not used by the Ministry as antenatal care providers.4 The clinics were 13 14 not able to carry out postnatal home visits as required by governmental standards.6 Poor 15 7 16 women were lessFor likely to havepeer postnatal review care. only 17 Poverty, deprived infrastructure, military checkpoints and armed Israeli settlers restrict 18 19 freedom of movement and reduces access to central health-facilities and legal assistance in 20 8,9 21 rural areas. During an escalation of the conflict between the years 2000 to 2006, it was 22 reported that 69 women gave birth at military checkpoints, causing casualties in both mothers 23 4 24 and babies, as they were not allowed to reach hospitals. Although the political situation in the 25 West Bank was less volatile in the following decade, rural women are still vulnerable and 26 27 depend more on governmental facilities than women in urban areas, as rural private services 28 9 29 are scarce. 30 Several studies describe how midwife-led continuity models improved health for mothers and 31 32 babies. Most studies were from high-income countries.10-16 The World Health Organisation http://bmjopen.bmj.com/ 33 34 (WHO) recommends implementation and research on midwife-led continuity models to 35 improve quality in low- and middle-income countries.17 Two main ways of organizing such 36 37 models are described in the literature. In the case-load model one midwife cares for up to 45 38 women and facilitate relational continuity, while in the team-midwifery model a group of 4-6 39 40 midwives can provide care for up to 360 women through the pregnancy, intrapartum and on September 29, 2021 by guest. Protected copyright. 41 42 postnatal period. Ideally, in both models, women during labour are cared for by a known 43 midwife.11,17 44 45 To improve services in rural areas, the Palestinian Ministry of Health, in cooperation with the 46 47 non-governmental humanitarian organization Norwegian Aid Committee (NORWAC), 48 launched the implementation of a modified midwife-led continuity, case-load-model, in 2013, 49 50 starting in the Nablus and Jericho Governmental hospitals and surrounding villages. The 51 implementation involved the communities as well as several levels in the Ministry of Health, 52 53 to overcome known barriers to quality of care.18 54 55 56 57 58 3 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 The aim of this study was to investigate whether the Palestinian midwife-led continuity model 4 had an impact on the utilization of maternity services and selected quality indicators at the 5 6 two regions’ clinics after two years of experience. 7 8 9 METHODS 10 11 Implementation of the Palestinian Midwife-led continuity model 12 The modified case-load model aimed at establishing a relationship between the pregnant 13 14 woman and her midwife, during pregnancy and the postnatal period. The midwife also 15 16 worked at the governmentalFor peer hospital, where review most women would only give birth unless they chose a 17 private hospital. 18 19 Once assigned, midwives in Nablus and Jericho governmental hospitals received training. 20 21 Under supervision, they provided antenatal care in clinics and postnatal home visits in the 22 surrounding villages. The hospital in Nablus had enough midwives to serve ten villages per 23 24 week, meaning two midwives would leave hospital each weekday, five days a week. 25 Midwives from Jericho hospital served five villages in the Jordan Valley, with one midwife 26 27 visiting one village every weekday. Three extra midwives were employed in the hospital in 28 29 Nablus, and two in Jericho, to maintain the capacity at the labour ward. All midwives worked 30 full time, as part time employment was not possible at the Ministry of Health. The same 31 32 midwife visited the same village, usually once a week. If the designated midwife was on http://bmjopen.bmj.com/ 33 34 holiday or sick leave one of the other midwives would cover her village. The ideal case load 35 per midwife was around 50 pregnant women yearly, but should not exceed 100. Thus, the 36 37 smallest village with 16 registered pregnant women per year was visited every second week 38 only and the largest village with 163 registered pregnant women was shared by two midwives 39 40 weekly. The remaining working days the midwife spent in the labour ward. All pregnant on September 29, 2021 by guest. Protected copyright. 41 42 women were informed that their midwife during pregnancy also worked at the local 43 governmental hospital. Independent of place of delivery, all women registered at the clinic 44 45 were to be offered postnatal home visits. All pregnant women were informed that the limited 46 47 numbers of midwives, and the large workload in the labour ward, made it difficult to ensure 48 they would meet the midwife they knew from antenatal care, during labour. Women were 49 50 given the phone number for their midwife in case of an emergency. The level of relational 51 continuity was limited to the antenatal and postnatal period. 52 53 Nevertheless, a relational continuity was possible also during labour, if their known midwife 54 55 happened to be on duty. Implementing the model aimed to strengthen the relationship 56 between the woman and her midwife, improve interdisciplinary cooperation, and reduce the 57 58 4 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 barrier between hospital and primary health care. The midwives received driving lessons to 4 obtain a driving license, and used designated cars with the Ministry of Health logo, and 5 6 marked Midwifery Care, to facilitate transportation to villages and homes. Standard care in 7 8 the clinics without this model was offered mainly by nurses or midwives, and medical doctors 9 (GP) working only in primary health care. 10 11 12 Study design 13 14 As the implementation started before the study, a non-randomized intervention design was 15 16 chosen to evaluateFor the model, peer based on registryreview data at cluster only level from two complete years 17 before (2011 & 2012), and two complete years after (2014 & 2015) the intervention. Clinics 18 19 where the model was implemented were compared with clinics where the model was not 20 21 implemented. Both arms of the study followed the same written governmental procedures. 22 The study was part of an implementation research project aiming at documenting the effect of 23 24 the midwife-led continuity model. The study was approved by the Regional ethical committee 25 of South East Norway and by the Palestinian Ministry of Health. 26 27 28 29 30 31 32 Clusters http://bmjopen.bmj.com/ 33 34 The clusters consisted of governmental primary health village-clinics in Nablus and Jericho 35 regions. There were 53 active clinics during the study period. During autumn 2013, the 36 37 midwifery model was implemented in 16 clinics. 38 All clinics were stratified for inclusion by rural and urban location, activity period, and 39 40 intervention period. on September 29, 2021 by guest. Protected copyright. 41 42 43 Exclusion criteria 44 45 Clinics located less than three km from Nablus and Jericho centre were defined as urban, thus 46 47 pregnant women had better access to private and non-governmental services. Clinics in urban 48 areas were therefor excluded. Clinics opened during the study period were excluded due to 49 50 incomplete data. 51 Clinics where the intervention was prematurely terminated or introduced later during the 52 53 study period, were excluded because of contaminated data. 54 55 56 Outcomes 57 58 5 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 The number of antenatal visits was chosen as the primary outcome. Secondary outcomes were 4 number of pregnant women referred to higher level of care, and number of women receiving 5 6 postnatal home visits. Other outcomes were number of women registered at the clinic for 7 8 antenatal care, number of pregnant women referred for abnormal blood sugar levels, number 9 of women seen by doctor after birth, number of newborns seen by doctor after birth, and 10 11 number of total postnatal consultations for mother and newborn. 12

13 14 Statistical analysis 15 16 Aggregated dataFor were retrieved peer from the reviewgovernmental registry. only The registry consisted of 17 anonymous data reported monthly from all clinics to the central statistical database in the 18 19 Ministry of Health. The registry did not include data at an individual level. 20 21 Mean, standard deviation and range were given for normally distributed and count variables. 22 Percentage and total number were given for categorical variables. 23 24 25 Change from baseline in the intervention and standard care groups, and any differences 26 27 between the groups changes, were examined by using mixed effects models. In the mixed 28 29 models, the clinic was specified as cluster (i.e. random variable), time and group, and 30 interaction between time and group were treated as fixed variables. 31 32 http://bmjopen.bmj.com/ 33 34 Approximately normally distributed count outcomes were fitted by mixed effects linear 35 regressions. Before fitting, the outcome variables were divided by the number of registered 36 37 pregnant women, or newborns, to calculate average values. Variance weights for each average 38 were then computed and included in a variance formula in the model because of 39 40 heteroscedasticity, since the computed averages are based on different numbers of pregnant on September 29, 2021 by guest. Protected copyright. 41 42 women. 43 44 45 Mixed effects Poisson regressions were used to fit non-normally distributed count outcomes, 46 47 and an offset variable was used to adjust for the total number of (individual) registrations that 48 were under risk in the models. 49 50 51 Measured confounding variables, which could have influenced the key estimates, were: the 52 53 village’s population size, whether the clinic had an employed community midwife and 54 55 laboratory equipment, regular military check points between village and hospital, and distance 56 from hospital. These possible confounders were included in the mixed models for adjusting. 57 58 6 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 4 Adjusted regression coefficient as means and rate ratios with 95% confidence intervals were 5 6 given. Two-sided p-values of less than 0.05 were considered statistically significant. 7 8 9 The map was developed by using ArcGIS software, and the attribute/database are part of the 10 11 same application. 12

13 14 The analyses of mixed effect models were performed with R version 3.4 and STATA version 15 16 14. Descriptive analysesFor were peer carried out review using IBM-SPSS versiononly 21 for Windows. 17 18 19 20 21 RESULTS 22 After stratification, 39 clinics were included in the study: 14 as intervention and 25 as control 23 24 clinics (Figure 1). 25 In total, 10 034 women booked at the 39 included clinics during the study period, 2 784 in the 26 27 intervention clinics and 7 250 in the control clinics. 28 29 The clinic locations are presented in Figure 2. 30 The Palestinian Ministry of Health confirmed that no other activities were introduced 31 32 unequally to the groups during the study period. The clinics were located in a region where http://bmjopen.bmj.com/ 33 34 political unrest and economic hardship most likely would affect the intervention and control 35 groups similarly during the study period. The measured possible confounders presented in 36 37 table 1 were adjusted for in the final results, none had significant confounding effect. 38 39 40 Table 1 Characteristics of clusters* on September 29, 2021 by guest. Protected copyright. 41 42 Characteristics Intervention Control 43 Population served mean 3402 4636 44 min/max 1000/7554 1875/11017 45 Distance to hospital (km) mean 23 12.6 46 min/max 5/59 3/28

47 Number of clinics with employed 0 8 48 community midwife 49 Number of clinics with laboratory 5 10 50 51 Additional clinics in village (NGO) 4 2 52 Number of clinic with regular military 6 14 53 checkpoints between village and hospital 54 * 14 clinics with intervention and 25 clinics with standard care(control) 55 56 57 58 7 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 4 Descriptive statistics of primary and secondary outcome variables such as number of 5 6 individuals registered, mean, standard deviation and range for the two timepoints in the 7 8 intervention and standard care are presented in table 2. 9 10 11 Table 2 Descriptive statistics for primary and secondary outcomes at baseline and two years after the 12 implementation Groups/clusters Mean 13 (n=39)* timepoint** N*** **** SD Range 14 min 15 max Before 1094 39 18 12 95 16 For peerIntervention review only 17 Number of women registered After 1690 60 33 16 163 18 for ANC during study period Before 3180 64 36 3 168 Standard care 19 After 4070 81 42 23 213 20 Total 10034 21 Before 2220 79 42 27 176 22 Intervention 23 Number of newborns registered After 2470 108 61 29 217 24 during study period Before 5416 88 46 33 291 Standard care 25 After 5771 115 67 36 298 26 total 15877 27 Before 4015 3.7 0.9 1.9 5.6 28 Number of antenatal visits and Intervention 29 mean visit per woman per After 7994 4.7 1.2 3.0 7.0 cluster (recurrent visits divided 30 Before 14657 4.6 1.4 2.4 8.6 on number of registered Standard care 31 women) 32 After 16769 4.2 1.1 2.4 6.8 http://bmjopen.bmj.com/ 33 Before 79 7.3 8.3 0 36.4 Number of referrals to higher Intervention 34 After 456 25.6 14.2 2.3 54.1 35 level of care (mean % of registered pregnant per cluster) Before 427 12.0 11.7 0 45.5 36 Standard care After 549 12.8 13.2 0 66.7 37 Before 12 1.7 4.7 0 17.7 38 Number of registered pregnant Intervention 39 women receiving home visits After 721 41.8 25.2 0 97.5 40 after birth (mean % of Before 42 1.5 4.9 0 25.5 on September 29, 2021 by guest. Protected copyright. 41 registered pregnant per cluster) Standard care 42 After 22 0.7 2.3 0 11.5 Before 57 27 21 109 43 Coverage - ratio % between of Intervention 44 number of registered newborns After 71 22 37 131 45 and registered pregnant per Before 61 23 7 119 46 cluster Standard care 47 After 74 16 47 113 48 Before 29 2.6 3.4 0 10.0 Number of referrals because of Intervention 49 abnormal blood sugar (mean % After 81 4.5 7.4 0 30.2 50 of registered pregnant per Before 90 2.6 3.5 0 14.9 51 cluster) Standard care After 105 2.7 3.6 0 13.0 52 53 Before 208 12.3 20.0 0 0.7 Number of mothers seen by Intervention After 461 20.1 32.4 0 94.0 54 doctor postnatally (mean % of 55 Before 534 12.1 20.5 0 93.0 registered newborns per cluster) Standard care 56 After 225 4.8 10.0 0 57.0 57 58 8 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Before 1670 79.6 24.5 29.7 118.6 Number of newborns seen by Intervention 4 After 2173 91.4 29.8 34.8 172.4 5 doctor postnatally (mean % of registered newborns per cluster) Before 4338 85.2 26.8 21.8 162.2 6 Standard care After 5082 90.4 21.6 47.5 142.1 7 Before 3902 1.8 0.3 1.3 2.2 8 Intervention 9 Total postnatal consultation for After 5364 2.2 0.4 1.7 3.1 10 newborn Before 9796 1.9 0.3 1.2 2.6 Standard care 11 After 10875 1.9 0.2 1.5 2.4 12 Before 1830 0.9 0.4 0.2 1.7 13 Total postnatal consultations Intervention After 3637 1.4 0.4 0.8 2.1 14 for mothers of registered Before 5073 1.0 0.5 0.2 3.1 15 newborn Standard care 16 For peer reviewAfter only5399 1.0 0.7 0.2 1.6 17 * 14 intervention clusters and 25 Standard care clusters** Two years before intervention (2011 & 2012) and two years 18 after intervention (2014 & 2015) *** Number of total individual registrations**** The mean at cluster level 19 20 21 Change within the intervention and standard care group as means and rate ratios (RRs), and 22 difference between the changes within the two groups, controlled for potential confounding 23 24 covariates, are presented in Table 3. 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 9 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 23

11 pregnant Control -0.39 (-0.73 - 0.05) p=0.026 12 13 Number of newborn’s Intervention 18.2% (10.0 - 26.4) p<0.0001 mothers who 14 6.6% (-3.1 - 16.4) 0.179 registered at clinic Control 15 during their pregnancy 11.6% (6.3 - 16.9) p<0.0001 16 For peer review only 17 Number of newborns Intervention 18 12.8% (-1.8 - 27.3) p=0.085 seen by doctor 4.7% (-12.7 - 21.9) 0.599 19 postnatally Control 8.1% (-1.3 - 17.5) p=0.089 20

21 Total number of Intervention 22 0.41 (0.26 - 0.57) p<0.0001 postnatal consultations 0.33 (0.16 - 0.52) 0.0004 23 for newborns Control 0.08 (-0.02 -0.18) p=0.126 24 25 Total number postnatal Intervention 0.64 (0.52 - 0.77) p<0.0001 26 consultations for 0.60 (0.46 - 0.75) <0.0001 27 mothers Control 0.04 (-0.04 -0.12) p=0.321 28 29 Adjusted RR (95% Outcome** Group P-value 30 Adjusted Rate Ratio RR (95%CI) CI) 31 32 Intervention 3.87 (3.04 - 4.92) p <0.0001 Number of referrals to http://bmjopen.bmj.com/ 3.64 (2.78 - 4.78) <0.0001 33 higher level of care Control 34 1.06 (0.94 - 1.21) p=0.353 35 Number of registered 36 Intervention 37.42 (21.14 - 66.22) p<0.0001 97.65 (45.20 - pregnant who received <0.0001 37 210.96) postnatal home visits Control 0.38 (0.23 - 0.64) p<0.0001 38

39 Number of referrals 40 Intervention 1.78 (1.16 - 2.72) p= 0.008 because of abnormal 1.83 (1.10 - 3.05) 0.021 on September 29, 2021 by guest. Protected copyright. 41 blood sugar Control 0.97 ((0.73-1.29) p=0.846 42 43 Number of mothers Intervention 1.94 (1.65 - 2.29) p<0.0001 44 seen by doctor 4.87 (3.88 - 6.10) <0.0001 45 postnatally Control 0.40 (0.34 - 0.47) p<0.0001 46 47 *Mixed effect linear regression was used to analyse change in and between clusters when data had normal 48 distribution**Mixed-effect Poisson regression was used to analyse the change in clusters when data was not normal 49 distributed8 ***All outputs were adjusted for potential confounders without any change in value. Covariates 50 adjusted for were: Distance from clinic to city hospital (km), population in village, if there were additional clinic in 51 village, clinics with community midwife, clinics with laboratory and clinics with a regular checkpoint on the way to 52 hospital. 53 54 55 56 Antenatal utilization 57 58 10 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 There was statistically significant difference in average change in mean number of antenatal 4 visits between the groups by 1.55 [1.38 – 1.54], p=0.0004. Mean number of visits increased 5 6 by 1.16 visits with the new model, while standard care declined with - 0.39 visits. In other 7 8 words, clinics with the new model had an increase from 3.7 mean number of antenatal visits 9 per pregnant woman before the model was introduced to 4.7 mean number of antenatal visits 10 11 per woman after, while in the control clinics, mean number of antenatal visits per woman 12 decreased from 4.6 to 4.2 visits. 13 14 15 16 For peer review only 17 Referrals 18 19 A statistically significant difference in change between the groups’ number of referrals to a 20 21 higher level of care was observed giving a ratio of rate-ratios of 3.64 [2.78 - 4.78], p<0.0001. 22 For the intervention group referrals increased by a rate-ratio of 3.87, meaning that the number 23 24 of referrals increased from 7.3% to 25.6% of all registered women in the clinics, while the 25 control group only had a change rate-ratio of 1.06, meaning that the percentage of referrals 26 27 moved only from 12 to 12.8% during the study period. 28 29 30 Postnatal service 31 32 Postnatal home visits increased substantially at the intervention clinics, whereas at the http://bmjopen.bmj.com/ 33 34 control-clinics it dropped giving a ratio of rate-ratios of 97.65 [45.20 – 210.96], p<0.0001. 35 With women in the intervention group, mean number of postnatal contacts with health 36 37 services increased significantly, whereas no increase was observed at clinics with standard 38 care, giving a ratio of rate ratios of 0.60 (95% CI 0.46 - 0.75) P<0.0001. Also, a significant 39 40 increase in mean number of newborn’s healthcare-contacts were observed with the new on September 29, 2021 by guest. Protected copyright. 41 42 model, but not in group with standard care, giving a ratio of rate ratios of 0.33 (0.16 - 0.52) 43 p=0.0004. 44 45 46 47 48 DISCUSSION 49 50 In the clinics with the midwife-led continuity intervention, a significant rise in mean number 51 of antenatal visits per woman was observed, whereas number of visits per woman decreased 52 53 in the clinics with standard care during the same period. It is thus likely that the improved 54 55 utilization was a result of the intervention. WHO recommended in 2002 a minimum of four 56 focused antenatal visits for healthy pregnant women.19 After evaluating new evidence, the 57 58 11 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 recommendation was revised in 2016 to a minimum of eight antenatal visits to reduce 4 perinatal mortality and improve women’s satisfaction.17,18,20 The women’s increased 5 6 adherence to service in clinics with midwife-led continuity may indicate that women 7 8 experienced improved quality of the services. Relational continuity is an important tool to 9 enhance communication and thus satisfaction with care.16 The association between improved 10 11 quality of care and increased utilization is supported by several previous studies, and by WHO 12 recommendations, to improve utilization and quality by introducing midwife-led continuity of 13 14 care.12,14,17,18 15 16 It is a possibilityFor that the pregnant peer woman review would feel safe knowing only that the midwife 17 following her throughout pregnancy also worked at the public hospital where she would give 18 19 birth, and that her midwife would visit her at home after birth. Due to the heavy workload, the 20 21 midwives could not be on call to attend birth. Midwife-led continuity of care in settings with 22 few midwife resources and heavy workload, must balance the demands on the midwives. 23 24 Although women were not guaranteed that the same midwife providing them antenatal and 25 postnatal care would attend their labour, their midwife’s connection to the governmental 26 27 hospital might have reduced the alien barrier to the hospital, and restored a feeling of security 28 29 for the rural women. 30 A qualitative study investigated midwives’ experience of working with a similar model in the 31 32 Ramallah region from 2007 to 2011. The midwives described how the model enabled them to http://bmjopen.bmj.com/ 33 34 give individualized care and how the broader scope of practice and increased autonomy gave 35 them important experience and tools for their work.21 This could serve to explain women’s 36 37 adherence to the antenatal service, because building a relationship with competent, respectful 38 and motivated midwives probably increased their wish to return to receive more care. 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 In the intervention group, a change in referral mechanisms was observed. The midwives 43 working with the new model identified significantly more risk factors leading to referral to 44 45 higher level of care, than in clinics with standard care. The finding of more women with 46 47 abnormal blood sugar level indicates that the model improved the identification of important 48 risk factors. The proportion of pregnant women referred to higher level of care increased to 49 50 25.6% with the new model vs. 13% with standard care during the study period. The 51 proportion in the new model is in line with the WHO-estimate presented in guidelines from 52 53 2001, that in general 25% of pregnant women would need additional antenatal care due to 54 22 55 health complications before or during pregnancy. The village clinics had little, if any, 56 57 58 12 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 technical resources to investigate risk signs, so referral to higher level of care was necessary 4 to follow up any possible complications. 5 6 One important quality indicator of antenatal care is the ability to detect possible complications 7 8 and involve specialist care when necessary. Kearns et al. (2015) demonstrate that improved 9 referral networks is a key element for improving quality in low-resource settings.23 The 10 11 process of information and referral within the system is also highlighted by the WHO 12 framework as one core indicator of quality of services.24,25 Some countries have much higher 13 14 level of maternal health risks than others due to poverty, high fertility rate and general health 15 10,26 16 challenges. AFor Palestinian peer study from review2015 revealed that 26.9%only of women who gave birth 17 experienced one or more morbidities.27 The rise in numbers of referrals after introducing the 18 19 new model matched the WHO and the local estimates. Thus, it seems reasonable to suggest 20 21 that the intervention improved the referral system. 22 23 24 The results furthermore showed a substantial increase in the number of postnatal care contacts 25 for mothers and new-borns, including home visits. The WHO recommendation for postnatal 26 27 care is a minimum of three postnatal contacts; and a minimum of one home visit preferably 28 28 29 during the first week after birth. The result from the study consequently indicates that the 30 implementation of midwife-led continuity models may contribute to reach such a goal. 31 32 Filby et al. (2016) describe how lack of transportation hamper quality improvements in other http://bmjopen.bmj.com/ 33 29 34 rural resource-constrained settings. The implementation in rural Palestine included a 35 designated vehicle and driving skills for the midwives, facilitation of transportation was 36 37 consequently a key factor in reaching out to the villages and home visits. 38 The fact that women receiving midwife-led continuity were more frequently seen by a doctor 39 40 in the clinic after birth (20.1% vs. 4.8%), in addition to the midwife, indicate improved on September 29, 2021 by guest. Protected copyright. 41 42 interdisciplinary cooperation. When midwives undertook home visits and discovered health 43 problems or risks, they involved the doctor. The findings also showed that there was a 44 45 systematic check of newborn babies by doctors in all the village clinics, and the home visit 46 47 from the midwife added to this. The increase in, and variation of, postnatal contacts including 48 home visits, make it reasonable to conclude that the midwifery model improved both 49 50 utilization and quality of postnatal care at a cluster level. 51 52 53 Limitations of the study 54 55 The study was carried out after the implementation of the midwife-led continuity model 56 started. This prevented a randomized allocation of clinics to intervention and control clusters. 57 58 13 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 The number of midwives available in the hospitals limited the number of clinics for 4 implementation to a total of 15 in these regions. The Ministry chose to implement the 5 6 program in the clinics they found appropriate. The baseline data show that a reason for 7 8 choosing these clinics were due to challenges in service provision, thus improvements could 9 have been easier achieved and lead to bias. 10 11 Another limitation was weak data because the facility-based registry did not include data at 12 individual level. This is a common problem in low-and middle income-countries.23 Lack of an 13 14 individual reproductive health registry prevented measuring individual impact and an intra- 15 16 cluster coefficient.For Thus, itpeer was not possible review to know, when womenonly registered at the clinic, or 17 who came back for recurrent antenatal-visits, or the reasons for referrals. 18 19 20 21 Strength of the study and further recommendations 22 The high number of clusters and the robust cluster data strengthen the study. The 23 24 organizational leadership, engagement and adherence of the multidisciplinary team strengthen 25 the sustainably of a complex intervention and its applicability to real life settings. The 26 27 findings make it reasonable to conclude that the new model had an effect on the utilization 28 29 and on some quality indicators of the maternal services. The study can be a useful tool in 30 power-calculations and planning of randomized trials for future implementation of the model. 31 32 http://bmjopen.bmj.com/ 33 34 Triangulation of methods within an implementation research framework would be useful to 35 investigate the broader effect of the implementation. This is highly recommended when 36 37 introducing evidence-based interventions to improve health service delivery in real world 38 settings where context is an important factor.30 Further research should be done to investigate 39 40 if the model could have impact on individual health outcomes for mother and newborn, and on September 29, 2021 by guest. Protected copyright. 41 42 on satisfaction with care. Previous research has described midwife-led models as a cost saving 43 way to improve maternal health in developing countries.12,31 A study of this model’s cost- 44 45 effectiveness would be useful. The general understaffing of both primary and secondary 46 47 governmental health services should be taken into consideration. This calls for an increase in 48 the number of midwives to improve quality. By implementing the model, more midwives 49 50 were employed at the hospitals to serve the community, enhancing the workforce of trained 51 midwives in both primary and secondary health service. The benefit of the midwife’s broader 52 53 scope of practice and experience, and the improved interdisciplinary cooperation should be 54 55 investigated. The cost of transportation was reduced to a minimum by enabling the midwives 56 57 58 14 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 to drive the vehicle themselves, as employing drivers would have added unsustainable cost to 4 the model. 5 6 7 8 Conclusion 9 The findings make it reasonable to conclude that the new model had effect on the utilization 10 11 and some quality indicators of the maternal service. 12 The positive change in facility-level outcomes show that clinics with the midwifery model in 13 14 the regions of Nablus and Jericho improved services during pregnancy and postnatally. The 15 16 findings indicateFor improvement peer of utilization review and some quality only indicators linked to facility- 17 level outcomes, such as continuity, functioning referral system and postnatal home visits. 18 19 The results of this study support the expansion of the model to new areas in Palestine. We 20 21 believe the model can be useful for other low- and middle-income countries to improve 22 utilization and quality of care. 23 24 25 26 27 Acknowledgements The authors would like to thank the Palestinian Ministry of Health for 28 29 their efforts in implementing the model, and preparing and sharing of their health registry. We 30 would also like to thank the Norwegian Aid Committee for supporting the study, and United 31 32 Nations, Office for the coordination of Humanitarian Affairs (OCHA) for developing the http://bmjopen.bmj.com/ 33 34 map. Finally, we are grateful to all the involved Palestinian midwives for their courageous 35 efforts in improving care for women in Palestine. 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 Contributors 43 BM was involved with the Implementation, study design, data collection, data analysis, data 44 45 interpretation and writing, MiL with study design, data interpretation and writing, LMD with 46 47 study design, data analysis and writing, MaL was involved with study design, data 48 interpretation and writing, AAA with the implementation, data interpretation and writing 49 50 MS conducted collection and systematization of data, EF involved in study design, data 51 collection, data analysis, data interpretation and writing. BM drafted the article, figures and 52 53 tables. All authors have reviewed and approved the final manuscript. 54 55 Funding 56 57 58 15 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 This work was partly supported by the Research Council of Norway through the Global 4 Health and Vaccination Program (GLOBVAC), project number 243706 and partly by public 5 6 funding through Norwegian Aid Committee. Three of the Authors were partly employed by 7 8 NORWAC and were involved in the implementation and interpretation of data. The analysis 9 was performed by a statistician at Oslo University hospital. The corresponding author had full 10 11 access to all the data in the study and had final responsibility for the decision to submit for 12 publication. 13 14 15 16 Competing interestsFor None peer declared review only 17 18 19 Ethics approval The study was approved by the Norwegian Regional Committee for Medical 20 21 Health research Ethics South East (REK) id number: 2015/1235. It was also approved by the 22 Palestinian Ministry of Health. 23 24 Data sharing statement The datafile is available upon request to the corresponding author 25 after receiving approval from the Palestinian Ministry of Health. 26 27 Figure legends: 28 29 Figure 1. Flowchart showing the stratification of all clusters in the study area. 30 31 Figure 2. Map showing the location of all included clusters in the study area. 32 http://bmjopen.bmj.com/ 33 34 35 36 37 1. World Bank website Country information, accessed September 2017 38 http://data.worldbank.org/country/west-bank-and-gaza 39 2. Palestine Ministry of Health Annual report 2013. Palestine,Nablus, Palestine Ministry 40 of Health, 2013. on September 29, 2021 by guest. Protected copyright. 41 3. Palestinian Central Bureau of Statistics, 2015. Palestinian Multiple Indicator Cluster 42 Survey 2014, Final Report, Ramallah, Palestine 43 4. Rahim HF, Wick L, Halileh S, et al. Maternal and child health in the occupied 44 Palestinian territory. Lancet 2009; 373(9667): 967-77. 45 5. Wick L, Mikki N, Giacaman R, Abdul-Rahim H. Childbirth in Palestine. Int J 46 47 Gynaecol Obstet 2005; 89(2): 174-8. 48 6. Abu Awad, Amal. PhD dissertation, Implementation of postnatal protocols by nurse in 49 primary health care clinics in the West Bank: University of Wisconsin-Madison 2011. 50 7. Nabaa HAA, Hilal GA, Sbeih SA, Ghandour R, Giacaman R. Access to care for 51 women reporting postnatal complications in the occupied Palestinian territory: a cross- 52 sectional study. The Lancet 2013; 382: S2. 53 8. Reem Al-Botmeh U. A review of Palestinian legislation from a women’s right 54 perspective. UNDP-papp-research-Legislative english.pdf UNDP, 2012. 55 9. United Nations, Office for the Coordination of Humanitarian Affairs occupied 56 Palestinian territory Area C Vulnerability Profile. http://data.ochaopt.org/vpp.aspx 57 58 16 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 23 BMJ Open

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Accessed September 2017. 4 10. Graham W, Woodd S, Byass P, et al. Diversity and divergence: the dynamic burden of 5 poor maternal health. Lancet 2016; 388(10056): 2164-75. 6 11. Homer C, Brodie P, Leap N. Midwifery continuity of care : a practical guide. Sydney; 7 New York: Churchill Livingstone/Elsevier; 2008. 8 12. Sandall J, Soltani H, Gates S, Shennan A, Devane D. Midwife-led continuity models 9 versus other models of care for childbearing women. Cochrane Database Syst Rev 10 2016; 4: CD004667. 11 13. ten Hoope-Bender P, de Bernis L, Campbell J, et al. Improvement of maternal and 12 newborn health through midwifery. Lancet 2014; 384(9949): 1226-35. 13 14 14. Renfrew MJ, McFadden A, Bastos MH, et al. Midwifery and quality care: findings 15 from a new evidence-informed framework for maternal and newborn care. Lancet 16 2014; 384For(9948): 1129-45.peer review only 17 15. Homer CS, Friberg IK, Dias MA, et al. The projected effect of scaling up midwifery. 18 Lancet 2014; 384(9948): 1146-57. 19 16. Homer CS. Challenging midwifery care, challenging midwives and challenging the 20 system. Women Birth 2006; 19(3): 79-83. 21 17. WHO. WHO recommendations on antenatal care for a positive pregnancy experience. 22 http://apps.who.int/iris/bitstream/10665/250796/1/9789241549912-eng.pdf?ua=1 23 World Health Organization; 2016. p. 152. 24 18. Nair M, Yoshida S, Lambrechts T, et al. Facilitators and barriers to quality of care in 25 maternal, newborn and child health: a global situational analysis through metareview. 26 27 Bmj Open 2014; 4(5): e004749. 28 19. Carroli G, Villar J, Piaggio G, et al. WHO systematic review of randomised controlled 29 trials of routine antenatal care. Lancet 2001; 357(9268): 1565-70. 30 20. Dowswell T, Carroli G, Duley L, et al. Alternative versus standard packages of 31 antenatal care for low-risk pregnancy. Cochrane Database Syst Rev 2015; (7): 32 CD000934. http://bmjopen.bmj.com/ 33 21. Mortensen B. To be weiled or not to be - what unite is the question, Experiences from 34 a continuity of Midwifery Care Model in Palestine and Norway. Bodø, Norway: Nord 35 University; 2011. p. 121. 36 22. Villar.J BP. Antenatal Care Randomized Trial: Manual for the Implementation of the 37 new model. . http://apps.who.int/iris/bitstream/10665/42513/1/WHO_RHR_01.30.pdf 38 UNDP, UNFPA, WHO, World Bank Special Programme of Research Development 39 and Research in Human Reproduction, 2002. 40 on September 29, 2021 by guest. Protected copyright. 41 23. Kearns AD, Caglia JM, ten Hoope-Bender P, Langer A. Antenatal and postnatal care: 42 a review of innovative models for improving availability, accessibility, acceptability 43 and quality of services in low-resource settings. Bjog-Int J Obstet Gy 2015; 123(4): 44 540-8. 45 24. World Health Organisation. Standards for improving quality of maternal and newborn 46 care in health facilities. WHO, Geneva, 2016 47 http://apps.who.int/iris/bitstream/10665/249155/1/9789241511216- 48 eng.pdf?ua=1 49 25. Tuncalp, Were WM, MacLennan C, et al. Quality of care for pregnant women and 50 newborns-the WHO vision. BJOG 2015; 122(8): 1045-9. 51 26. Oyibo PG, Ebeigbe PN, Nwonwu EU. Assessment of the risk status of pregnant 52 women presenting for antenatal care in a rural health facility in Ebonyi State, South 53 54 Eastern Nigeria. N Am J Med Sci 2011; 3(9): 424-7. 55 27. Hassan SJ, Wick L, DeJong J. A glance into the hidden burden of maternal morbidity 56 and patterns of management in a Palestinian governmental referral hospital. Women 57 58 17 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 23

BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from 1 2 3 Birth 2015; 28(4): e148-56. 4 28. World Health Organisation. Recommendations of postnatal care for mothers and new 5 born, 2013 The Departments of Maternal, Newborn, Child and Adolescent Health and 6 Reproductive Health and Research of the World Health Organization. 7 http://apps.who.int/iris/bitstream/10665/97603/1/9789241506649_eng.pdf?ua=1 8 World Health organization; 2013. p. 62. 9 29. Filby A, McConville F, Portela A. What Prevents Quality Midwifery Care? A 10 Systematic Mapping of Barriers in Low and Middle Income Countries from the 11 Provider Perspective. PLoS One 2016; 11(5): e0153391. 12 30. Peters DH, Adam T, Alonge O, Agyepong IA, Tran N. Implementation research: what 13 14 it is and how to do it. BMJ 2013; 347: f6753. 15 31. Friedman HS, Liang M, Banks JL. Measuring the cost-effectiveness of midwife-led 16 versus physician-ledFor peer intrapartum reviewteams in developing only countries. Womens Health 17 (Lond) 2015; 11(4): 553-64. 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 18 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 23 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Figure 1. Flowchart showing the stratification of all clusters in the study area. http://bmjopen.bmj.com/ 33 34 174x136mm (300 x 300 DPI) 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 23 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 Figure 2. Map showing the location of all included clusters in the study area. 46 47 209x297mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 23 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

1 TREND Statement Checklist 2 3 Paper Item Descriptor Reported? 4 Section/ No 5 Pg # 6 Topic 7 Title and Abstract 8 9 Title and 1 Information on how unit were allocated to interventions 1 10 Abstract Structured abstract recommended 1 11 Information on target population or study sample 1 12 13 Introduction 14 Background 2 Scientific background and explanation of rationale 1 & 2 15 Theories used in designing behavioral interventions 16 For peer review only 2 17 Methods 18 19 Participants 3 Eligibility criteria for participants, including criteria at different levels in 3 & 4 20 recruitment/sampling plan (e.g., cities, clinics, subjects) 21 Method of recruitment (e.g., referral, self-selection), including the 22 sampling method if a systematic sampling plan was implemented 3 & 4 23 Recruitment setting 3 & 4 24 Settings and locations where the data were collected 3 & 4 25 Interventions 4 Details of the interventions intended for each study condition and how 26 3 27 and when they were actually administered, specifically including: 28 o Content: what was given? 3 29 o Delivery method: how was the content given? 3 30 o Unit of delivery: how were the subjects grouped during delivery? 31 o Deliverer: who delivered the intervention? 32 3 http://bmjopen.bmj.com/ 33 o Setting: where was the intervention delivered? 3 34 o Exposure quantity and duration: how many sessions or episodes or 35 events were intended to be delivered? How long were they 3 36 intended to last? 37 o Time span: how long was it intended to take to deliver the 38 intervention to each unit? 3 39 o Activities to increase compliance or adherence (e.g., incentives) 40 3 on September 29, 2021 by guest. Protected copyright. 41 Objectives 5 Specific objectives and hypotheses 3 42 Outcomes 6 Clearly defined primary and secondary outcome measures 4 43 Methods used to collect data and any methods used to enhance the 44 quality of measurements 4 45 Information on validated instruments such as psychometric and biometric 46 4 47 properties Sample Size 7 How sample size was determined and, when applicable, explanation of any 48 4 49 interim analyses and stopping rules 50 Assignment 8 Unit of assignment (the unit being assigned to study condition, e.g., 51 Method individual, group, community) 4 52 Method used to assign units to study conditions, including details of any 53 4 54 restriction (e.g., blocking, stratification, minimization) Inclusion of aspects employed to help minimize potential bias induced due 55 5 56 to non-randomization (e.g., matching) 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 23 BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from

51 Baseline comparisons of those lost to follow-up and those retained, overall 52 and by study condition Comparison between study population at baseline and target population 53 7 & 8 54 of interest 55 Baseline 15 Data on study group equivalence at baseline and statistical methods used 56 equivalence to control for baseline differences 8 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 23 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2017-019568 on 22 March 2018. Downloaded from