The content and completeness of women-held maternity documents before admission for labour: A mixed methods study in Banjul, The Gambia.

BACKGROUND: Women-held maternity documents are well established for enabling continuity of maternity care worldwide, with the World Health Organisation (WHO) recommending their use in effective decision-making. We aimed to assess the presence, content and completeness of women-held maternity documents at admission to hospitals in The Gambia, and investigate barriers and facilitators to their completion.
METHODS: We interviewed 250 women on maternity wards of all 3 Banjul hospitals and conducted content analysis of documentation brought by women on admission for their completeness against WHO referrals criteria. Logistic regression models were used to estimate the odds of the minimum criteria being met. Two focus groups and 21 semi-structured interviews (8 doctors, 8 midwives and 5 nurses) were conducted with healthcare practitioners to explore barriers and facilitators to documented clinical information availability on admission.
FINDINGS: Of the women admitted, all but 10/250 (4%) brought either a maternity card or a structured referral sheet. Of all forms of documentation, women most frequently brought the government-issued maternity card (235/250, 94%); 16% of cards had all 9 minimum criteria completed. Of the 79 referred women, 60% carried standardised referral forms. Only 30% of 97 high-risk women had risk-status recorded. Women were less likely to have documents complete if they were illiterate, had not attended three maternity appointments, or lived more than one hour from hospital. During qualitative interviews, three themes were identified: women as agents for transporting information and documents (e.g. remembering to bring maternity cards); role of individual healthcare professionals' actions (e.g. legibility of handwriting); system and organisational culture (e.g. standardised referral guidelines).
CONCLUSION: Women rarely forgot their maternity card, but documents brought at admission were frequently incomplete. This is a missed opportunity to enhance handover and quality of care, especially for high-risk women. National guidelines were recognised by providers as needed for good document keeping and would enhance the women-held maternity documents' contribution to improving both safety and continuity of care.

Background

Reducing maternal mortality is a high priority on the international health agenda[1]. In 2013, The Gambia had a maternal mortality ratio (MMR) of 433 per 100,000 live births, one of the highest globally[2,3]. Attempts to reduce maternal mortality in line with the Sustainable Development Goals are challenging in this resource-limited setting; delays or errors to decision-making processes due to an inadequate maternity history or case documentation contribute to these maternal deaths[4-6]. The World Health Organisation(WHO)’s 2016 ‘Standards for Improving Quality of Maternal and Newborn Care in Health Facilities’ emphasises that increasing maternal health facility coverage is not enough and that improving the quality of information and referral systems, specifically documentation, is one of the keys to improving outcomes[7].

Women-held documents are well established in maternity care worldwide; at least 163 countries are known to use some form of home-based record[8-11]. Based on research, the WHO continue to emphasise their use to help in decision-making and ‘continuity, quality of care and [a mother’s] pregnancy experience’[11-16]. Additionally there is good evidence that women-held documents have other advantages, including that they help women to feel empowered and more involved in their own health and that of their babies[16-19].

Women-held documents often take the form of maternity cards. The maternity card has the potential to assist in continuity of care within maternity services as it enables the handover of clinical information, between antenatal appointments, at admission to maternity units and during post-natal care[11,14,19]. As women frequently move from one facility to another during their pregnancy in both low and high-income countries, women-held documents can ensure the clinical history is available and so pregnancy complications are more likely to be detected and acted upon[14,16,18,19,20,21]. A WHO collaborative study in 1993 specifically highlighted the potential use of maternity cards to enhance the diagnosis and referral of high-risk women[16]. However, to fulfil this function, the cards must be consistently brought to appointments by women, looked at by staff and filled out in their entirety with contents that fulfil WHO recommendations.

Whilst the maternity card is recommended, its feasibility and use in a real-world (especially resource limited) setting is important to establish. The maternity card is a low-cost resource already widely accepted across low-income countries (LICs)[11,20] and research has suggested that women rarely forget their documents at antenatal appointments[16,22,23]. However, systematic reviews state that no studies have assessed the availability of antenatal records at the time of delivery (rather than between antenatal appointments)[22] or reported on the content or completeness of maternity women-held documents at the time of birth[18,19]. This is arguably the most important time when antenatal handover information is required.

Establishing the effectiveness and use of maternity card use in The Gambia may yield findings relevant to other LICs and countries in the region that face similar challenges to maternal health. This study further addresses research gaps laid out by both the WHO and recent systematic reviews[11,18,19].

The primary aim of this study was to assess the number, type, content quality and completeness of women-held documents on admission to maternity units in The Gambia, a LIC with a high MMR[2]. Secondly, we aimed to explore context-specific barriers and facilitators to effective use of women-held documents in maternity units by health professionals and maternity staff, especially for women admitted with high-risk pregnancies or deliveries.

Methods

This was a convergent parallel mixed-methods study[24] that took place across all three maternity hospital departments in the Greater Banjul and Kanifing region between January and March 2018 (S1 Table). Antenatal services are well attended in The Gambia with over 90% visiting a clinic at least once during their pregnancy at a variety of health facilities ranging from mobile health posts and local health centres to the tertiary hospital in Banjul[25] The government-issued yellow maternal card is an A4 double-sided piece of card supplied on the first antenatal appointment and intended to stay with the woman until her final post-natal check-up (Figs 1 and 2).

Fig 1

The front page of the government issued women-held ‘yellow Maternity Card’.

Risk status was sometimes added in red ink at the top of the card. A midwife has written a warning about document completeness on this particular card.

Fig 2

The inside of the government issued women-held ‘yellow Maternity Card’.

Generally this side of the card contains information relevant post-discharge.

Quantitative data

In-patient women on antenatal, postnatal and maternity high-dependency units, aged 16 and over at the three study hospitals were invited to participate whilst waiting to leave the ward after formal ‘discharge’. Researchers rotated around the three hospitals across an even distribution of days throughout the study period (including weekends) and between each facility to increase the likelihood of obtaining a representative sample during the study period and to reduce any differences between observers. Researchers were present throughout the period of the day that discharges took place (between 9am and 1pm) and recruited all eligible women discharged that day who gave informed consent (thumbprint or signature). Women taking part in any Medical Research Council (MRC) study (n = 21; to comply with local MRC ethical approval requirements) and women unable to speak English, Mandinka, Wolof or Fula (n = 0) were excluded. Ethical approval was granted by the Scientific Coordinating Committee, the joint Government/MRC Ethics Committee in The Gambia and the University of Birmingham BMedSc Population Sciences and Humanities Internal Research Ethics Committee.

To achieve a 95% confidence level (α = 0.05) with a ±5% accuracy, a minimum sample size of 243 was required to estimate the number of women who brought a maternity card, based on a population of an unknown size[26]. This was assuming that an 80% prevalence of women would bring their maternity card to the maternity unit, based on our unpublished research in Kerala, India[27].

A verbal questionnaire, adapted and piloted from studies in Mongolia[28] and India[29], was administered on wards with the help of trained local interpreters (S1 Text). The researcher also conducted a document review by recording type and contents of any documentation brought to or received from the ward (including the maternity card) and ward-based patient medical records were reviewed to establish the reason for women’s admission.

Content quality and completeness was determined by comparison with minimum criteria from WHO recommendations for maternity referrals that focus on emergency situations (Fig 3)[30] and analysed as median number of criteria fulfilled (criteria fulfilled: yes or no). Although not all admissions were under emergency situations, these criteria were deemed appropriate on consultation with UK/Gambian maternal health experts, including the UK Royal College of Obstetrics and Gynaecology. Additional items considered of interest and those that had a designated space on the maternity card were assessed but not included in the minimum criteria analysis.

Fig 3

Minimum criteria.

Minimum criteria for referral pattern improvement were adapted from WHO criteria[30] after consultation with maternal health experts at the UK Royal College of Obstetrics and Gynaecology. Criteria are applied to all admissions on the inference that the transfer of care from antenatal services to maternity unit admission is a form of ‘referral’.

Data was analysed using SPSS 24.0 (IBM, Armonk, NY, USA). Descriptive statistics (numbers and percentages) characterised the nature and quality of documents; including the number of each type of document, whether individual criteria were met and how many women carried documents that met the minimum criteria. Subgroups of women who were referred, attended scanning, or were high-risk were explored. Percentages were calculated using non-missing data as the denominator.

To establish whether any particular characteristics could predict whether the minimum criteria would or would not be met by a woman’s documents, binomial logistic regression models were used to produce both unadjusted and adjusted odds ratios. Predictor variables were entered into the model based on clinical rationale and on findings from the qualitative arm of the study. The rationale was supported by evidence from Mongolia and India that low socioeconomic status and living far from the hospital can lower the quality of their written documentation[11,28,29]. Occupation was regarded as the best representation of socio-economic status (preferable to house structure as most women lived in cement and corrugate property). Therefore; age, occupation, time to get to hospital, number of children, English literacy, whether they had been referred, and number of previous contacts were all deemed to be potentially associated with document completeness. Address, risk status and transport to hospital were initially included but were removed from the model due to multi-collinearity in association with other variables, as revealed through examining the correlation matrices and cross-tabulations. ‘Hospital 1, 2 or 3’ was also added to the regression model as a fixed variable to adjust for any hospital effects.

Qualitative data

In order to establish barriers and facilitators to effective use of maternity cards and documentation, qualitative and quantitative components of the study were conducted in parallel in all 3 hospitals. A purposive sampling frame included all cadres of healthcare professionals (HCPs) including nurses, midwives and doctors (see S1 Table for numbers of staff who work at each hospital). Initially, we planned semi-structured interviews (SSIs) to obtain in-depth and sensitive views and experiences which may be difficult to elicit in a group setting and focus group discussions (FGDs) to explore more in-depth group dynamics, agreed behaviours and ways of operation. However, the FGDs proved difficult to arrange due to staff availability; it was seemingly rare for large numbers of staff to be available at the same time. For the SSIs, sampling was much more systematic. Participants were allocated to either FGDs or one-to-one SSIs; no staff participated in both. Participants were recruited until thematic saturation was achieved. FDGs and SSIs were conducted in English (all HCPs spoke fluent English) at the hospital sites using a pre-determined topic guide (S2 Text) that addressed views on current practices of information exchange, barriers and facilitators to effective handover using the documents, and feasible opportunities for improvement. The FDGs and SSIs were recorded, then transcribed verbatim and anonymised by the researcher.

Inductive thematic analysis based on Braun and Clarke’s six-step approach[31] was undertaken to identify themes from the data. The researcher performed line-by-line coding on all transcripts and another researcher independently coded four of the most data-rich transcripts for analyst triangulation[32]. Themes and subthemes were subsequently developed and refined. Convergent triangulation was used to combine quantitative and qualitative results in the discussion.

Due to resource and time constraints, it was not possible to conduct qualitative research on the women’s perspectives of the maternity card process for handover by HCPs. The authors collected qualitative data concerning the major barriers or facilitators to the effective completion and use of the card by HCPs rather than women’s views of health-care providers use of the card.

Results

Quantitative component results

In total, 251 eligible women were approached. None refused, but one lacked capacity to consent hence was not included. 250 women completed the questionnaire in the study period. This represented approximately 25% of all discharges from the three facilities throughout the study period (total 1,082; Hospital 1–193, Hospital 2–371, Hospital 3–518). Table 1 shows women’s demographic and admission characteristics.

Table 1

Women’s demographic and admission characteristics across the three hospitals study sites.

Background Characteristic	Hospital 1 (n = 72)	Hospital 2 (n = 91)	Hospital 3 (n = 87)	Total (n = 250)
	No. (%)	No. (%)	No. (%)	No. (%)
Age*
20 and under	11 (15.3)	16 (17.6)	17 (19.5)	44 (17.6)
21–29	31 (43.1)	38 (41.8)	45 (51.7)	114 *45.6)
30 and over	30 (41.7)	37 (40.7)	25 (28.7)	92 (36.8)
Occupation*
Housewife	48 (66.7)	52 (57.1)	62 (71.3)	162 (64.8)
Retail	12 (16.7)	26 (28.6)	11 (12.6)	49 (19.6)
Othera	23 (16.7)	13 (14.3)	14 (16.1)	39 (15.6)
Time taken to get to hospital*
Under 1 hour	43 (61.4)	58 (63.7)	86 (98.9)	187 (75.4)
1 hour and above	27 (38.6)	33 (36.3)	1 (1.1)	61 (24.6)
Transport used to get to hospital
Walked	5 (6.9)	1 (1.1)	8 (9.2)	14 (5.6)
Taxi / Gelli	31 (43.1)	76 (83.5)	79 (90.8)	186 (74.4)
Ambulance	36 (50.0)	14 (15.4)	0 (0.0)	50 (20.0)
Number of childrenb *
0	3 (4.2)	6 (6.6)	3 (3.4)	12 (4.8)
1–2	36 (50.0)	44 (48.4)	47 (54.0)	127 (50.8)
3–4	18 (25.0)	19 (20.9)	21 (24.1)	58 (23.2)
5–6	11 (15.3)	9 (9.9)	10 (11.5)	30 (12.0)
More than 6	4 (5.6)	13 (14.3)	6 (6.9)	23 (9.2)
English literacy*
Yes	39 (54.2)	46 (50.5)	57 (65.5)	142 (56.8)
No	33 (45.8)	45 (49.5)	30 (34.5)	108 (43.2)
Education
None/incomplete primary	17 (23.6)	23 (25.3)	11 (12.6)	51 (20.4)
Primary/Secondary	30 (41.7)	35 (38.5)	46 (52.9)	111 (44.4)
Higher	3 (4.2)	3 (3.3)	5 (5.7)	11 (4.4)
Islamic or other	22 (30.6)	30 (33.0)	25 (28.7)	77 (30.8)
Address
Combo/Banjul/Kanifing	37 (51.4)	35 (38.5)	82 (94.3)	154 (61.6)
West Coast	20 (27.8)	55 (60.4)	5 (5.7)	80 (32.0)
Provinces/’Up-river’	15 (20.8)	1 (1.1)	0 (0.0)	16 (6.4)
Structure of house
Brick and tiles	9 (12.5)	8 (8.8)	3 (3.4)	20 (8.0)
Mud/sand and corrugate	14 (19.4)	14 (15.4)	3 (3.4)	31 (12.4)
Cement and corrugate	49 (68.1)	69 (75.8)	81 (93.1)	199 (79.6)
Tribe
Mandinka	24 (33.3)	37 (40.7)	33 (37.9)	94 (27.6)
Fula	18 (25.0)	17 (18.7)	20 (23.0)	55 (22.0)
Wolof	13 (18.1)	15 (16.5)	20 (23.0)	48 (19.2)
Other	17 (23.6)	22 (24.2)	14 (16.1)	53 (21.2)
Admission Characteristic
Referred*
No	22 (30.6)	54 (59.3)	84 (96.6)	160 (64.0)
Yes	50 (69.4)	37 (40.7)	3 (3.4)	90 (36.0)
High riskc
Yes	47 (65.3)	45 (50.6)	11 (12.6)	103 (41.5)
No	25 (34.7)	44 (49.4)	76 (87.4)	145 (58.5)
Missing		2d		2d
Number of previous contacts*
1 to 3	21 (29.6)	30 (33.3)	27 (31.0)	78 (31.5)
More than 3	50 (70.4)	60 (66.7)	60 (69.0)	170 (68.5)
Missing	1d	1d		2d

* Denotes that the variable was entered into the regression analysis.

a. Other occupations include: farmer, student, tailor, civil servant (e.g. police).

b. Number of children in mothers’ family, not including current pregnancy/baby born on that admission.

c. High-risk is defined as; multi-pregnancy, pre-eclampsia or pregnancy induced hypertension, severe anaemia, previous C-section/forceps/ventoux delivery and past medical history of diabetes or heart condition (and age <14 years but not applicable). High parity was also considered a risk factor by some of the staff but this was not consistent or featured in guidelines and so was not included (S3 Text).

d. Number of missing data points where information was not available. Valid percentages have been calculated from available information.

On arrival at the maternity department, documents brought by women included: maternity cards (Figs 1 and 2), structured referral sheets (Fig 4), and a selection of less frequently presented documents (ultrasound reports, prescription notes, scraps of paper, child health reports, miscellaneous lab requests/results). All but 10/250 (4%) of women had brought either a structured referral sheet or a maternity card.

Fig 4

Structured referral sheet.

Each hospital (and many of the health centres that referred to the hospitals) had centre-specific sheets designed for external referral. All had similar section headings, but documents were not standardised by the government (unlike the maternity card).

Maternity cards

The standard government issue yellow women-held maternity card (Figs 1 and 2) was brought by 94.0% (235/250, 90.3–96.3%) of women and a further 1.2% (3/250, 0.4–3.5%) brought an alternative maternity card e.g. from a private clinic. Of the 238 cards, the content of 2 could not be assessed. However, 80.1% (189/236) were incomplete with at least one unfilled feature; 26.7% (63/236) noted the ‘Estimated Date of Delivery’ [Table 2]. The maternity card was brought by 94.2% (97/103) of high-risk women, but only 29.9% (29/97) had their status recorded as high-risk on their card (rose to 36.4% [36/97] when ‘other document’ contents were included) [Table 2]. When risk-status was available on the card, it was normally part of the obstetric history section or written on the top of the card in red (Figs 1 and 2) as there was no designated space for this information. Although accuracy could not be assessed for all fields, evident inaccuracies were also noted; for example, 3% (7/236) of maternity cards had the wrong age recorded based on mothers’ reports to our data collectors.

Table 2

Contents of women-held documents at admission.

Document Content Item	Any Document	Maternal Card	Referral Sheet	Other documents a
	N = 250b	N = 236c	N = 44d	N = 99
	No. (% of cases)	No. (% of cards)	No. (% of sheets)	No. (% of ‘other’s)
MINIMUM CRITERIA
1. Mother’s name	242 (96.8)	236 (100)	44 (100)	97 (98.0)
2. Age	241 (96.4)	235e (99.6)	34 (77.3)	74 (74.7)
3. Address	238 (95.2)	231 (97.9)	39 (88.6)	51 (51.5)
4. Parity	231 (92.4)	230 (97.5)	4 (9.1)	19 (19.2)
5. Estimated day of delivery	88 (35.2)	63 (26.7)	1 (2.3)	34 (34.3)
6. Complications in antenatal period f	241 (96.4)	209 (88.6)	9 (20.5)	31 (31.3)
7. Relevant past obstetric complications g	214 (85.6)	209 (88.6)	17 (38.6)	3 (3.0)
8. Treatments/tests applied thus far	217 (86.8)	202 (85.6)	24 (54.5)	96 (97.0)
9. Results of treatment/tests	194 (77.6)	165 (69.9)	21 (47.7)	92 (92.9)
Problem referred for	60 (24.0)	20 (8.5)	41 (93.2)	5 (5.1)
Recommended place of delivery	97 (38.8)	97 (41.1)	2 (4.5)	65 (65.7)
Gravida	214 (85.6)	213 (90.3)	6 (13.6)	15 (15.2)
HIV status	4 (1.6)	2 (0.8)	0 (0.0)	2 (2.0)
Emergency/risk status	52 (20.8)	38 (16.1)	17 (38.6)	2 (2.0)
Medications	242 (96.8)	212 (89.0)	23 (52.3)	14 (14.1)
Contraception	239 (95.6)	200 (84.7)	2 (4.5)	0 (0.0)
Detail is illegible	37 (14.8)	29 (12.3)	6 (13.6)	11 (11.1)
Appears incomplete (at least 1 doc)	199 (79.6)	189 (80.1)	20 (45.5)	28 (28.3)
- 2 documents	33 (13.2)
- 3 documents	2 (0.8)

a. Combination of all other documents (including ultrasound reports, lab requests, child health booklets, discharge cards, prescription notes and miscellaneous)

b. Percentages were calculated using non-missing data as the denominator (N)

c. 2 antenatal cards were brought by women but unable to be assessed for content and completeness

d. 4 referral sheets were brought by women but were unable to be assessed for content and completeness

e. 7 maternity cards had the wrong age recorded (3%)

f. ‘Complications in antenatal period’ was regarded as completed if there was any information regarding antenatal history. On the maternity card, this would mean it should always have at least a single entry of the antenatal check-up where the woman had been issued with the card.

g. ‘Past obstetric complications’ was regarded as any information regarding obstetric history. On the maternity card, this could be left empty if the woman was primiparous. However, it was regular practice for staff to have written ‘n/a’ or ‘primi’ to indicate this, which we took to be desirable practice.

Referral sheets

Referrals (36%, 79/250) were generally emergency cases when a standardised referral sheet, issued by the Ministry of Health, was expected to be brought in addition to the maternity card: 59.5% (47/79) of referred women had a structured referral sheet (Table 2). 81.1% (9/11) of those referred from another hospital carried a referral sheet, compared with only 58.6% (34/58) from a health centre. Of the referral sheets brought, the content of 3 could not be assessed. None of the referral sheets carried all the minimum criteria for safe maternal handover, although 93.2% (41/44) included the ‘reason for referral’. Despite a referral sheet providing a place to mark if the case was emergency/non-emergency, only 38.6% (17/44) had this information completed.

Ultrasound scanning and estimated date of delivery

Most (82%, 205/250) women had attended at least one scan to provide accurate estimate of the date of delivery. Results were routinely recorded on an ultrasound report sheet, not the maternity card. However, only 58.8% (100/170) of women who reported attending scanning brought an ultrasound scan report to hospital (35 women who had attended scanning were not able to be assessed for ultrasound scan presence in their documentation due to a researcher error in data collection).

Minimum criteria fulfilment

When all individual documents were combined, 24.4% (61/250) of women brought documentation that overall met all nine minimum WHO referral criteria. The median score was 8 out of 9 (IQR 7–9) and 68% (170/250) of women had at least 8 of the criteria fulfilled. Estimated delivery date was the least well-fulfilled criteria (64.8% of respondents lacked it; 162/250). Of all documents, maternity cards had the highest median criteria score and were the only document that provided all 9 minimum criteria, however only 15.7% (37/236) of the cards achieved this (S1 Fig).

Minimum criteria scores were categorised into scores below 8 (insufficient, 32.1% of respondents, 80/250) or 8 and above (sufficient 67.9%, 170/250). ‘Sufficient’ was not defined as 9/9 criteria in the analysis because of limited numbers meeting all 9 criteria (61/250). In logistic regression analysis, being literate in English (OR 2.04 [95% C.I. 1.08–3.85]), having 1–4 children compared to having fewer or more (OR 4.4 [95% C.I. 1.04–18.07]), having more than 3 contacts with healthcare during pregnancy (OR 2.16 [95% C.I. 1.15–4.03]) were all positively and significantly associated with minimum criteria fulfilment. Travelling further than 1 hour to get to hospital (OR 0.34 [95% C.I. 0.15–0.74]) and attending a hospital other than the tertiary referral centre (Hospital 2 OR 0.45 [95% C.I. 0.19–1.02], Hospital 3 OR 0.22 [95% C.I. 0.08–0.60]) were negatively associated with minimum criteria fulfilment (Table 3).

Table 3

Results of logistic regression analyses exploring associations between women’s characteristics and the likelihood of their documentation containing at least 8 out of 9 minimum criteria.

Independent Variable/Characteristic	Unadjusted		Adjusted
	OR (95% CI)	P-value	OR (95% CI)	P-value
Age of mother
Under 20	1a	0.31	1a	0.052
21–29	0.61 (0.27–1.37)		0.28 (0.10–0.78)
30 and over	0.53 (0.23–1.20)		0.36 (0.12–1.09)
Occupation of mother
Housewife	1a	0.516	1a	0.098
Retail	0.79 (0.40–1.56)		0.68 (0.31–1.47)
Other	0.67 (0.33–1.34)		0.40 (0.17–0.94)
Time taken to get to hospital
Under 1 hour	1a	0.039	1a	0.007
1 hour and above	0.53 (0.29–0.97)		0.34 (0.15–0.74)
Number of children
0	1a	0.046	1a	0.13
1–5	3.58 (1.09–11.79)		4.40 (1.04–18.07)
5 or more	2.13 (0.60–7.62)		3.81 (0.76–19.06)
English literacyb
Illiterate	1a	0.22	1a	0.029
Literate	1.88 (1.10–3.21)		2.04 (1.08–3.85)
Referred for careb
No	1a	0.429	1a	0.741
Yes	1.25 (0.71–2.20)		0.88 (0.40–1.93)
Number of contacts with healthcare throughout pregnancy
3 or less	1a	0.004	1a	0.016
More than 3	2.26 (1.29–3.96)		2.16 (1.15–4.03)
Hospital
1	1a	0.11	1a	0.013
2	0.53 (0.26–1.06)	0.074	0.45 (0.19–1.02)
3	0.49 (0.24–1.00)	0.048	0.22 (0.08–0.60)

a. Categories of predictor variables that received ORs of 1.00 are reference categories

b. Correlation matrices revealed strong multi-collinearity between ‘referred for care’, ‘brought by ambulance’, and ‘high-risk’. Therefore referral was selected as the most appropriate variable for the model as it is the most likely of the three to have an impact on the document type and completeness. ‘English literacy’ and ‘education’ were also associated and literacy was selected for the same reason.

Qualitative results

SSIs were conducted with 21 members of the multidisciplinary teams in all hospitals (8 doctors, 8 midwives and 5 nurses). Two FGDs were conducted, one consisting of five midwives and the other consisting of four nurses, in one of the three hospitals. The demographic details of the participants interviewed are shown in S2 Table.

Three themes describing facilitators and barriers to effective handover at admission using women-held documents were identified: women as agents for transporting information and documents; role of individual healthcare professionals’ actions; system and organisational culture. These themes are presented in Table 4 along with sub-themes and supporting quotations.

Table 4

Qualitative results support quotations of barriers and facilitators to effective handover on admission through use of documentation.

Theme	Barrier/facilitator	Sub-theme	Quotations
• Women as agents for transporting information and documents	Facilitator	• Women normally bring documents	“most of our patients, er when they come back for deliveries, their delivery or any admission or come in for any visits, they come back with their antenatal card, previous investigations like the blood investigations, urine and ultrasound scanning” (D5)
	Barrier	• Losing smaller sheets e.g. prescription notes so information not in the card is not available	“Sometimes the patient will miss their prescription because it’s a small sheet, they misplace it and they cannot report to you what medications they take” (D5)
Role of individual HCPs for provision of information and documents	Facilitator	• If everything is written clearly on the maternity card, it can aid handover	“If everything is written on the [antenatal] card, the treatment and the date of the visit and not only on the prescription part, then it will help a lot.” (D5)
		• Women not able to understand the medical terms (low health literacy) making written notes more important	“Sometimes the patients, we will ask them but not all of them are able to speak. Not all of them are able to say or understand the medical terms, the medical issues but from the paper we can cross-check and say oh these things have happened” (N2)
		• Improvisation of HCPs for highlighting high-risk patients in documentation	“from the clinic, they will just put high-risk on the [maternity] card and why. Some will be high parity, some will be pre-eclampsia, high BP.” (M3)
	Barrier	• Not enough information written on referral sheet	• “you don’t have enough information in that referring sheet… Most of the time… the referring doctors or referring nurses that are in other health facilities don’t write enough information.” (D1)• “if it’s a referral, they now bring their antenatal card… sometimes the documentations are not enough [information], most times they’re not enough.” (D7)• “the referral notes, mostly they don’t have enough information. Sometimes, they don’t have the contact of the referring officer… Sometimes what-all that is done for the patient mostly are not there, sometimes even the vitals sometimes they miss it.” (N5)
		• Lack of clarity	“some people will just put high-risk-hospital delivery [on the maternity card] but they will not say why.” (M3)
		• Illegible handwriting	• “Most of the time, you need to clerk again because the referring doctors or referring nurses that are in other health facilities don’t write enough information, or some don’t even write legibly for you to be reading” (D1)• “their handwriting is bad . . . Handwriting is important because you’re writing for someone to read so if you’re writing it and someone else can’t read it so it’s useless, it’s like don’t write.” (D7)
		• Inaccurate written information for referral	“for example weekends er Fridays are mostly the days that we have er-most of the days that we sometimes have a lot of referrals because they want to empty their hospitals so they can enjoy their weekend… sometimes even it’s-they write things that are not even happening” (D7)
		• Lost information on small pieces of paper	“every investigation is attached to the [maternity] card… sometimes we will miss it because it’s a small sheet, they [patients] misplace it. That is where the deficit comes.” (D5)
System and organisational culture	Facilitator	• Standardised referral forms	“the country developed a referral form document which if you are referring a patient for every health centres… Even the private clinics, they too… have their own referral forms.” (M1)
		• Designated space on referral forms for feedback	“in the referring form, there is a place where feedback should be given to the previous centre, but like it’s not done.” (N7)
		• Communication between referring and receiving health centre facilitated by use of structured cards	“the referral form… it has to be filled accordingly… the way you received the patient, what you managed with the patient and why you want to refer it here, you understand, so that you can avoid unnecessary referrals” (D1)
	Barrier	• Lack of supervision and reinforcement of structured referral forms	• “there is a national problem in regards to the communication from the referral centre to the receiving centre…you are just here sitting or somewhere busy doing other things and then a patient arrive in a very critical situation…” (D3)• “you might not even know the number of the referral centre” (M8)• “Sometimes, they [the referral notes] don’t have the contact of the referring officer.” (N5)
		• Lack of referral guidelines	“I think still now there is a challenge on that, from the referral centre to the receiving centre… we need like a protocol or guideline.” (D3)
		• Poor attitude of staff and organisational culture on filling referral forms and maternity cards	“everything that you have done for the patient you have to write feedback [on the referral form]… some may be lazy to do the documentation part for the feedback.” (M6)
		• Unqualified staff accompanying women being referred to the receiving facility	• “the ones doing the referrals are not the ones bringing the patient… they assign a very junior nurse… so they when they come, they just give you the referral form. When you ask, they say I don’t know anything.” [M12]• “the referrals are left with untrained nurses” [M11]
		• Non-antenatal patients never issued with documents and so poor handover information available	• “'Where is your antenatal card' 'I've never gone to an ANC' . . . when you see that you know have a big challenge to do.” (M8)• “sometimes you know they [the women] come and they have no medical papers.” (N2)• “They [the women] normally have [antenatal] cards… but some will stay at home without that, you will just, you will see them on their day of delivery . . . you don’t know nothing” (M4)

The code letter after each quotation refers to the cadre of HCP; doctor (D), midwife (M) and nurse (N).

The first theme groups the barriers and facilitators of women as agents for transporting information and documents. Participants reported that whilst women do normally bring documents (itself a facilitator), some individuals arrive with no documents. One doctor explained that this is often due to the woman never having attended an antenatal appointment before. The general view was that when they do bring the documents, the documents are extremely important for care as they facilitated handing over the antenatal information. This was particularly significant since many of the patients cannot explain the medical issues themselves, commonly due to lack of education and health illiteracy. Respondents also described how smaller loose sheets inside the booklet, such as prescription notes, were often misplaced by the women. Missing notes can result in difficulty in deciding on clinical pathways for the healthcare staff upon admission to the hospital, especially if the women themselves are incapable of explaining their health issues.

The second theme considers the role of individual HCPs’ actions as barriers and facilitators to effective handover. Participants suggested that when all the information is written on the antenatal card or the referral form rather than on separate pieces of paper as notes inside the card, this ‘will help a lot’ on arrival for admission. Incomplete documentation resulted in valuable time being wasted as the admitting healthcare staff would often have to conduct extra thorough examinations and guess the emergency cause of referral which can lead to errors, whereas in the presence of a good history or good handover notes, immediate management could begin. Multiple participants highlighted illegible handwriting as a key barrier to documentation use. Lack of detail and clarity about the exact reason for a woman being “high-risk” was flagged as an issue. Moreover, individual HCPs at health-centre facilities were reported to be referring women inappropriately to ease pressure on their own facilities. This led to inaccurate reasons for referral on the official referral forms in order to justify the referral, which could have detrimental impacts on patient care and safety.

The third theme recognises how system and organisational facilitators and barriers exist with the use of documents on admission. Contributing factors to individual staff inappropriately referring women include both staff shortages and resource scarcity in their own facilities. For high-risk patient transfers, this also led to reliance on unqualified staff accompanying the woman being referred in an ambulance to the receiving facility not being able to explain the problems of the transferred women. This placed greater importance on the content of documents carried by the woman, including her maternity card and her referral form. Some participants suggested that an electronic notes based system connecting all centres could facilitate better use of medical documentation through increased information exchange efficiency between healthcare teams and overcome the issue of illegible handwritten notes.

Discussion

Both the quantitative and qualitative arms of this study found that the majority of women delivering in the 3 Banjul and Kanifing hospitals were successfully issued with a maternity card during pregnancy that they brought with them to hospital. However, many of the cards and the referral sheets were incomplete and neither regularly met the WHO minimum content criteria for referral; including that one third of women who were high-risk did not have their risk status recorded. Both the card and the referral sheet were often ‘inaccurate’ and loose sheets tucked inside the card were often lost. The number of contacts with healthcare during pregnancy, distance from the hospital, and a woman’s literacy all influence the completeness of her documents.

Our LMIC sample of pregnant women performed similarly–if not better–than high income countries (HICs) on bringing their women-held documents to hospital. For example, in Australia studies and audits have shown 85–93% compared to the 94% of women in the Gambia[33,34]. This supplements previous literature describing women-held document use between antenatal appointments and agrees that women reliably bring their maternity cards[16,22,23].

However, with regards to other documentation carried with the maternity card and completeness of the documents themselves, the situation was not so positive. It was found that the majority of cards were incomplete and HCPs complained of facing difficulties when admitting a woman as a result. Neither the combination of all documents brought by women, nor the maternity card itself, regularly met all the minimum criteria recommended by WHO and other documents only increased the proportion of women bringing minimum criteria by around 10%. To the best of our knowledge this is the first study to explore the completeness of women-held maternity documents in a LIC[18,19] although the omission of essential clinical information hindering efficient and safe healthcare delivery has been highlighted in previous studies exploring health information exchange in low and middle-income countries[35,36].

To improve patient safety and prevent maternal deaths, it is critical to have available on admission clinical information for high-risk women to prevent delays for essential decision-making and interventions. It is also important for post-natal care to know if the pregnancy had complications or was high-risk. Yet, only 2 in 3 high-risk women had their risk status available in their documentation, and only half of those had it on their maternity card. When risk status was recorded, HCPs stated that it was often unclear why that woman was high-risk. For the high-risk referred women, the incomplete maternity cards became more critical since only half of women referred to hospitals, generally high-risk admissions, carried a referral sheet with the maternity card. These forms were equally incomplete and HCPs described them as “frequently inaccurate”. If these women are also less health literate (unable to “understand the medical terms”), HCPs explained that documentation was more heavily relied upon as they cannot explain their own conditions. As such, we would recommend a new designated section on the maternity cards where high-risk reason could be marked.

Qualitative data confirmed the quantitative evidence of inaccuracies and losses of scan and test results on small pieces of paper usually slipped inside of the card. A further designated space for test results on the maternity card might help overcome this. The WHO’s 2018 “Evidence review of home-based maternal records” key-informant data suggested that documents could often be incomplete and inaccurate due to HCPs’ views of documentation as an unnecessary task (“double-work”) or having illegible handwriting[11]. Similarly, in our study we found participants explaining that HCPs in The Gambia may be ‘lazy’ with regards to documentation and that illegible handwriting on documents can cause ‘time wasting’ when trying to clerk the patient on admission.

Qualitative data from HCPs specifically noted that more guidelines and protocols are needed regarding use of referral sheets and filling of maternity cards, which echoes the WHO evidence review suggestion from key informants[11]. If staff could be given training to follow defined guidelines and supervised instructions to write all essential patient data on the same document (e.g. the maternity card), information may be more consistently available. Training could also help motivate behaviour change by explaining the benefits of good documentation and how maternity cards are perceived to be helpful by clinicians.

Significant differences existed between the completeness of documents at each of the hospitals and women who travelled further to get to the hospital had less complete documents. A study of patient held health booklets in Mongolia (for NCDs) had a similar finding[28]. Improved national referral procedures and better national level guidelines for the use of documentation, as suggested by HCPs in this study, could overcome this difference and may help standardise the completeness of documents.

Characteristics of women that were associated with increased completeness of documents included having had more than three antenatal visits; presumably as more frequent reviews offer more opportunity to complete documents. Trials in Thailand, Indonesia and Cambodia have all shown that use of a maternity card is associated with increased antenatal attendance[37-39]. Furthermore, women who were literate were more likely to have complete documentation, as was seen in and investigation of patient-held health booklets in Mongolia[28]. This is potentially a reflection of socioeconomic status, whereby more literate women are likely to attend facilities, demand and receive better care. Therefore, care must be taken to ensure that illiterate women do not continue to receive less complete documents and potentially perpetuate inequalities, since these women are often those with the highest risk of poor outcomes[12].

Limitations of the study included possible recall bias as women were interviewed about their admission at the time of discharge. Misplaced documentation during women’s admissions may have underestimated the number of documents, like ultrasound reports, that may have been brought with them during admission but lost during the inpatient period. The ‘minimum content’ completion of the admission documents may have been over-estimated, as hospital staff may have filled in certain admission sections of the maternal card whilst the woman was on the ward. Since urban hospitals were sampled, the results may be less generalisable to rural Gambia than the urban Banjul area, although all rural areas around Banjul did refer women to these three hospitals. In Pakistan it was shown that maternity cards were more effective in rural than urban centres[40]. On discussion with public health colleagues in the Gambia and reflecting on our study results that show women further from the hospitals had less complete documents, we would expect document use and completeness might be lower. This may be due to differing resource levels at the rural hospitals and clinics that might place more time-pressure on clinicians.

The strength of this study was that by employing mixed-methods, it enabled us to provide a more complete and comprehensive commentary. Not only have we been able to comment on the completeness of documents, but we have also been able to go some way to explain reasons for why documents were not complete. We included all maternity hospitals in the capital city of Banjul and thus did not have any sampling biases at the level of institutions.

Future studies could include rural areas and referring primary health centre facilities to increase generalisability and to fully understand the perspective of both ends of the referral system. Longitudinal studies that consider maternal outcomes relationship with document completeness would be challenging but important to investigate the importance of continuity of care to patient outcomes and to promote quality improvement interventions. Further research could also investigate opinions of the women themselves about the use of women-held documentation by HCPs in maternity care in LMIC.

Conclusion

We found that in The Gambia’s capital city Banjul and surrounding Kanifing region maternity hospitals, mothers universally carried their maternity cards and all health care providers referred to them. While the recommendations behind women-held documents in maternity services are clear[11-16], the finding that simple low-cost steps to improve the information recorded on the documents could be important to consider for this and other resource limited settings. Simple adaptations to the maternity card (such as spaces for test results and risk-status) and their better completion could capitalise on their almost universal use by women and staff to improve continuity of care and safer births. This is particularly important for high-risk deliveries if referral forms and other documentation continue to be absent or inadequately completed. Similar standardisation of referral forms, alongside national-level guidelines, training, supervision or monitoring for staff may ensure effective completion and maximise use of the cards and referral forms. Together, this would ensure all essential information is available to provide the smoothest handover to hospital-based care for births and prevent any delays to effective treatment and management of complications.

With WHO standards now emphasising quality of maternity care in hospitals, good handover of clinical information to ensure patient safety is likely to improve maternal outcomes, since handover of information is the cornerstone of patient safety and quality of care services[41]. With further development, women-held documents have the potential to play a greater role in the effective information transfer and referral systems in LICs and could optimise delivery of care and the reduction of global maternal mortality.

Chart showing total and individual document completeness against WHO minimum criteria for referrals.

(TIF)

Click here for additional data file.

Table of hospital background information.

Antenatal care is provided as part of the Maternal, Child health and Family Planning program (MCHFP) and can take place at a variety of health facilities ranging from mobile health posts and local health centres to the tertiary hospital in Banjul. Some health centres have birth facilities, others only provide antenatal care and tell mothers to go to hospital to deliver. Primary healthcare centres can refer women to any of the three hospitals, normally the closest maternity unit geographically. Women experiencing complications in hospitals in provinces further inland (‘upcountry’) are sometimes referred to Hospitals 1 or 2.

(DOCX)

Click here for additional data file.

Table of qualitative participant demographics information.

(DOCX)

Click here for additional data file.

Table of reasons for admission according to woman’s own description.

Answers were not mutually exclusive, respondents could select more than one option—therefore %s do not sum to 100. Answers are as patient described, unprompted.

(DOCX)

Click here for additional data file.

Self-designed questionnaire.

(DOCX)

Click here for additional data file.

Interview topic guide for FGDs and SSIs.

(DOCX)

Click here for additional data file.

Definitions of ‘High-Risk’ and ‘Complications’.

(DOCX)

Click here for additional data file.

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

4 Aug 2019

PONE-D-19-17140

Is the women-held maternity card enabling continuity of care on admission to maternity services in low-income settings? A mixed methods study in Banjul, The Gambia

PLOS ONE

Dear Senior Clinical Lecturer Manaseki-Holland,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We advice you to pay special interest in the comments of reviewer 3 and reviewer 4.

We would appreciate receiving your revised manuscript by Sep 18 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Astrid M. Kamperman

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

1. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

Reviewer #3: No

Reviewer #4: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: No

Reviewer #4: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

Reviewer #4: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: this study is well written and very specific to people in field. one of few research that merge quantitaive and qualitative data.

Reviewer #2: This mixed methods study on improving continuity of maternity care by using maternity card. The specific design was convergent parallel mixed-methods study, of which the quantitative component was the interview questionnaire survey and the qualitative component comprised of 2 focus groups and 21 in-depth interviews of healthcare professionals.

Isn’t the women-held maternity documents already well established? If so, the study should focus on either the feasibility in a local setting or completeness/accuracy of the records but not the efficacy in general. I think the authors already took these points into account as reflected by the design, results, and discussion. Hence, some of the texts (e.g. Line 59-61), and the title, should be revised at the authors’ discretion.

The qualitative results were too brief and did not integrate the quantitative findings as proposed. The strength of the mixed methods design was not utilized as anticipated.

Reviewer #3: Review Comments to the Author:

The topic of maternity cards is an interesting and important one and I applaud the authors for their work. I strongly support mixed methods manuscripts; however, this one tries to cover too much ground, and unfortunately, the quality of the text and analyses were sacrificed as a result. I have concerns about the quality of this study and the manuscript itself, which I describe in detail below.

I suggest that the coauthors regroup and decide on a narrower focus of a potential future manuscript. As a side note, in a future manuscript, I suggest incorporating language around how maternity cards can be empowering to women in taking care of themselves and advocating for their care.

Is the manuscript technically sound, and do the data support the conclusions?

The background section does not adequately cover what is already known about the use of maternity cards. The authors describe this as an efficacy study, but it more closely aligns with an effectiveness study since it takes place in a real-world setting.

The authors do not mention the sampling procedure for women participating in the quantitative proportion of the study. Did researchers interview each woman who was discharged until they reached a certain number of women on that day? On line 162, the authors write that research days were selected “at random and including weekends.” The authors fail to describe the specific procedure for randomization. Ultimately, the methods section needs to be written so that these findings could be reproduced.

Further, on line 204, the authors mention that participant were assigned to FGDs or IDIs solely based on their availability. This raises a concern as these two qualitative methodologies have very distinct and differing purposes. For example, FGDs are often used to understand group norms or group dynamics while an IDI allows for a private, in-depth conversation. Additionally, I would want to know why women were not interviewed to understand their perspectives of the maternity card process. If it was a matter of resources, that is fine, but it would have been good to acknowledge that in the text.

There are too many tables - with much more data than what is covered in the text. There are often not adequate text descriptions of the data in the tables. The discussion and conclusion sections are broad and would need to go deeper with the implications of the research as well as recommendations.

Has the statistical analysis been performed appropriately and rigorously?

Statistical analyses have not been performed appropriately and rigorously. In the presentation of descriptive statistics, there were often differing denominators and it was not always clear why. Confidence intervals were erroneously included when presenting descriptive statistics. For multivariate analyses, adjusted odds ratios were not included. Additionally, the authors do not control for clustering of the hospitals.

Is the manuscript presented in an intelligible fashion and written in standard English?

The language in the manuscript is not always intelligible or well organized. Additionally, the presentation of numbers and percentages are not consistent, which is quite distracting.

Reviewer #4: This is a very well-written paper describing interesting, and alarming, results of a mixed-methods study regarding woman-held documents in maternity care of the Gambia. The abundance of data is logically organized and intelligible, and results highlight specific areas for improvement in the completion of cards and, in turn, potential reductions in maternal mortality for the region. The authors conclusions are sound and suggestions for inclusion of space for risk-status and expected delivery date on maternity cards are fitting of the data and very reasonable recommendations for improving maternal care. However, there are aspects of the methods and discussion that detract from the current impact of this paper. Specifically, the methods used to develop the model require more detail to support their proper interpretation. A clearer description of why variables were included in the regression models is needed. Additionally, the discussion should be expanded to better explain and situate the findings. While this may be the first study to assess completeness of cards, the authors should enrich the discussion with comparisons of their results on rates of women carrying documents and qualitative findings with research in other LMIC.

Specific comments:

Line 96: It is unclear why the authors refer to anecdotal evidence?? Is there not empirically-based data? A reference to anecdotal data does not belong in the abstract.

Line 123: LICs should be spelled out first (i.e. “low-income countries (LICs)”). Line 136 should be changed to just LIC.

Line 127: Missing a “)” after (10).

Figure 1: If possible, please rotate the top image so it is easier to read.

Lines 193-197: The method by which variables were chosen to be included in regression analyses needs to be made clearer. Why were these specific few variables chosen? For example, what does distance from hospital have to do with completion of cards? What were the “appropriate exploratory models”??

Table 2 should have asterisk, like Table 1, indicating which variables were included in the modeling.

Line 247: Shouldn’t this be 79/250 not 251?

Line 334-335: Please expand on the studies referenced here. Did they have similar findings to the present study? Why yes or no?

Lines 354-367: What about the other significant sociodemographic factors? What hypothesis do the authors have for the association between sociodemographics (being a housewife, living close to the hospital, literacy) and card completion?

Lines 361-362: This appears to be the only reference to the qualitative data in the discussion. The authors should expand on how their findings are related to previous research and/or describe more the relation between their quant and qual data.

Line 362: I believe this should be changed to “in our sample, 100%..”. “In fact” seems to indicate that staff always review 100%, which cannot be inferred from this data.

Line 381: Please expand on in what ways you would anticipate the situation to be different in rural Gambia and why.

Line 387: What type of sub-group analyses were under-powered? What analysis of high-risk women was underpowered? Almost half of the sample was high-risk.

Line 392: Mother should be changed to mothers.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Dr Mohamad Al-Tannir

Reviewer #2: Yes: Krit Pongpirul

Reviewer #3: No

Reviewer #4: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

9 Nov 2019

1. This study is well written and very specific to people in field. one of few research that merge quantitative and qualitative data.

Thank you, we are glad that the benefits of merging quantitative and qualitative data are seen.

Maternity care is a high priority for the world health organization and sustainable development goals (as part of SDG 3). Although the sub-topic of maternity cards is specific, they are widely used, especially in low-middle income countries. Furthermore, they have featured in WHO guidelines both in the past and as recent as 2018, despite most of the evidence for them being ‘low-certainty’. Maternity cards are the cornerstone of antenatal care and delivery in Low-Middle income countries due to their facilitation of handover (which is key to patient safety across all specialties and countries). We hope that this study, although specific to the Gambia and maternal health, will be relevant to other low-income countries (LICs) and countries in the region that face similar challenges, especially as there is a limited amount of evidence regarding maternity cards from LICs. The effectiveness and feasibility of the card and some of the problems we identified could be generalised to a much wider number of countries where such research is yet to be done.

2. Isn’t the women-held maternity documents already well established? If so, the study should focus on either the feasibility in a local setting or completeness/accuracy of the records but not the efficacy in general. I think the authors already took these points into account as reflected by the design, results, and discussion. Hence, some of the texts (e.g. Line 59-61), and the title, should be revised at the authors’ discretion.

We agree with this comment that we have not conducted an efficacy study. Therefore have made changes in the wording in our ‘aims’ and throughout the paper to emphasise that the study explores the completeness of documents in a real-world setting. We have also made changes to the title to make this clearer.

3. The qualitative results were too brief and did not integrate the quantitative findings as proposed. The strength of the mixed methods design was not utilized as anticipated.

We understand that we have missed an opportunity to really make the most of the mixed-methods design. Therefore we have made multiple changes.

1) The qualitative analysis was reviewed and more supporting quotations and subthemes were added to the qualitative results table

2) Qualitative results that support the quantitative findings were more clearly signposted in the discussion. Furthermore, some of the qualitative findings were relevant in explaining the results of the rerun of the multivariate analysis of the quantitative data.

3) Qualitative results were used to help select clinically important variables for the rerun of the logistic regression which has been noted in the methodology section

4) Qualitative literature found in a new literature search and from the WHO 2018 Evidence Review of maternity documents was added to the background.

4. The topic of maternity cards is an interesting and important one and I applaud the authors for their work. I strongly support mixed methods manuscripts; however, this one tries to cover too much ground, and unfortunately, the quality of the text and analyses were sacrificed as a result. I have concerns about the quality of this study and the manuscript itself, which I describe in detail below.

I suggest that the coauthors regroup and decide on a narrower focus of a potential future manuscript. As a side note, in a future manuscript, I suggest incorporating language around how maternity cards can be empowering to women in taking care of themselves and advocating for their care.

Thank you very much in acknowledging the importance of this topic. As mentioned above in response to comment 3, we realise we have not made the most of the mixed-methods design and so have worked on multiple ways of integrating the findings (see above).

We recognise that much of the ‘too much ground’ partly stems from lack of integration of the quantitative and qualitative findings. It was also a product of finding that the situation on the ground was far more complex than simply the use of the maternity card, and synthesis of the results regarding ‘other documentation’ and ‘referral sheets’ presented challenges. Through significant reorganisation of tables, renewed analysis and added evidence to the qualitative findings, we have tried to substantially strengthen the foundations of our results.

Specifically, Table 1 and 2 have been combined and reorganized into the three hospitals rather than referred/not referred. We have converted one of the tables into a visual chart. We have moved the table describing reasons for admission to supplementary materials. We have removed the table discussing ‘missing data’ which confused the interpretation of the results. We have rerun the logistic regression analysis and revised the table accordingly.

Throughout the revised manuscript we have tried to put focus and emphasis on the fact that the paper is about the completeness of the documents as for the documents to have the desired improved effects on care, they have to be complete.

Our background section now tries to acknowledge that there is a good existing evidence base for women themselves appreciating the maternity card for the ‘empowerment’ and involvement in their own care. A previous and renewed literature search revealed that the major gap in research was partly on the feasibility of the card and its completeness in a (especially low-income) setting. We have added an explanation that resource and time restraints meant that it was not possible to conduct interviews with women, which might have revealed themes regarding empowerment that were not identified through just HCP interviews. We recognise that the women have an important opinion to voice, but this would be an important future research possibility (and was actually a feature in our concurrent research regarding discharge documentation). The combination of the good existing evidence base, lack of time and resources, and desire to keep the study focused on completeness of the documents, meant that we were not able to incorporate the impact of maternity cards on the women’s sense of empowerment.

5. The background section does not adequately cover what is already known about the use of maternity cards. The authors describe this as an efficacy study, but it more closely aligns with an effectiveness study since it takes place in a real-world setting.

We agree that this is certainly not an efficacy study. We have changed the title and references to the study type throughout the manuscript accordingly.

We have expanded our background section to include parts of our literature search, and multiple references, that we were previously required to remove due to work count limits and limits to the number of references allowed for our original submission to PLOS Medicine. Therefore, we hope the background section now adequately covers what is known about the maternity cards. Notably, since the original literature search, a ‘WHO Review of Evidence’ has been published regarding women-held documents in maternity care and this has been commented upon heavily in the new background section and discussion.

6. The authors do not mention the sampling procedure for women participating in the quantitative proportion of the study. Did researchers interview each woman who was discharged until they reached a certain number of women on that day? On line 162, the authors write that research days were selected “at random and including weekends.” The authors fail to describe the specific procedure for randomization. Ultimately, the methods section needs to be written so that these findings could be reproduced.

We have made appropriate changes to our methodology section to describe more accurately our sampling procedure.

We attempted to purposively select an even distribution of weekdays and weekends across the three study sites to increase the likelihood of obtaining a representative sample. We then recruited as many women being discharged on that day as possible. We did so by being present on the ward from 9 until 1 as we had information from staff that that was the time period that discharges took place. We then asked the women to wait for us to speak to them before they left. Unfortunately we do not know exactly how many women were missed each day, although we can comment on the total number of discharged and that we approximately sampled a quarter of all discharges from the three hospitals in the period

7. Further, on line 204, the authors mention that participant were assigned to FGDs or IDIs solely based on their availability. This raises a concern as these two qualitative methodologies have very distinct and differing purposes. For example, FGDs are often used to understand group norms or group dynamics while an IDI allows for a private, in-depth conversation.

Initially, we planned semi-structured interviews (SSIs) to obtain in-depth views and experiences which may be difficult to elicit in a group setting and focus group discussions (FGDs) to explore more in-depth group dynamics, agreed behaviours and ways of operation. However, the FGDs proved difficult to arrange due to staff availability; it was seemingly rare for large numbers of staff to be available at the same time. For the SSIs, sampling was much more systematic. Participants were allocated to either FGDs or one-to-one SSIs; no staff participated in both.

^This above paragraph has been included in the qualitative methodology section.

8. Additionally, I would want to know why women were not interviewed to understand their perspectives of the maternity card process. If it was a matter of resources, that is fine, but it would have been good to acknowledge that in the text.

As has been mentioned above (response 6), because the study objectives were mainly around the feasibility and completeness of the documents, it was considered that it would be more valuable to interview HCPs about the use of the cards and potential reasons why the cards would not be completed accurately. Only HCPs write in the cards and so completeness depends on the HCP not the woman. Additionally, time and resource constraints played a part in the decision. We have added a paragraph to explain this decision in the qualitative methodology section and have added mention of how this would be important for future study in the discussion.

9. There are too many tables - with much more data than what is covered in the text. There are often not adequate text descriptions of the data in the tables.

The discussion and conclusion sections are broad and would need to go deeper with the implications of the research as well as recommendations

The tables have been significantly restructured and more description of the tables has been added as footnotes.

Specifically, Table 1 and 2 have been combined and reorganized into the three hospitals rather than referred/not referred. We have converted one of the tables into a visual chart. We have moved the table describing reasons for admission to supplementary materials. We have removed the table discussing ‘missing data’ which confused the interpretation of the results. We have rerun the logistic regression analysis and revised the table accordingly.

The discussion section has been expanded to include more of the qualitative findings and to more explicitly explore the findings of the logistic regression analysis. We have tried to emphasise the clear recommendations of the study – to make simple changes to the existing card, standardise the referral documents, include all essential information on the maternity card, and introduce national level guidelines – that we feel are best supported by our evidence.

10. Statistical analyses have not been performed appropriately and rigorously. In the presentation of descriptive statistics, there were often differing denominators and it was not always clear why. Confidence intervals were erroneously included when presenting descriptive statistics. For multivariate analyses, adjusted odds ratios were not included. Additionally, the authors do not control for clustering of the hospitals.

We have sought further expert statistical advice from two statisticians and efforts have been made to review the presentation of our results. The denominators should now be clearer, and if they have changed due to missing data points, this is now explicitly mentioned in the text or footnotes of the tables.

Confidence intervals have been removed from the presentation of descriptive statistics, as there inclusion was indeed erroneous.

The results of both unadjusted and adjusted odds ratios from the logistic regression analysis are presented in table 3. The way in which variables were selected for inclusion in the logistic regression has been carefully described in the re-written methods section.

As there were only three study sites, ‘hospital’ was included as a fixed predictor variable in the final model to control for its influence.

11. The language in the manuscript is not always intelligible or well organized. Additionally, the presentation of numbers and percentages are not consistent, which is quite distracting.

Alongside significant restructuring of the tables and the addition of a bar-chart, we hope that the presentation of our data is clearer. We have added descriptions and footnotes to the tables to explain the data in a clearer manner. Significant changes to the text in all sections of our paper have been made and we have tried to make improvements to the clarity of language used. As we have removed the erroneous presentation of confidence intervals in the descriptive data, and have reviewed the denominators, we hope that the numbers are now clearer. We now consistently present a % and n/N for every piece of data stated in the results section.

12. This is a very well-written paper describing interesting, and alarming, results of a mixed-methods study regarding woman-held documents in maternity care of the Gambia. The abundance of data is logically organized and intelligible, and results highlight specific areas for improvement in the completion of cards and, in turn, potential reductions in maternal mortality for the region. The authors conclusions are sound and suggestions for inclusion of space for risk-status and expected delivery date on maternity cards are fitting of the data and very reasonable recommendations for improving maternal care. However, there are aspects of the methods and discussion that detract from the current impact of this paper. Specifically, the methods used to develop the model require more detail to support their proper interpretation. A clearer description of why variables were included in the regression models is needed.

We have sought further expert statistical advice and the regression analysis has been re-run. Characteristics, in the new analysis, were chosen based on clinical rationale, supported by both literature and the findings from the qualitative arm of the study. If variables were found to be too closely related on examination of cross-tabulations and the correlation matrices resulting in multi-collinearity, then they were excluded and clinical rationale was used to choose the most appropriate variable. This process has been explained in the methods section.

The results of the new regression analysis have been explored far more extensively in the discussion and we linked them to the qualitative findings and the literature to better understand these associations.

13. Additionally, the discussion should be expanded to better explain and situate the findings. While this may be the first study to assess completeness of cards, the authors should enrich the discussion with comparisons of their results on rates of women carrying documents and qualitative findings with research in other LMIC.

We have added clearer reference to relevant studies into women carrying documents in the second paragraph of the discussion (and we hope throughout). Other literature has also been used to help understand the new regression analysis results.

14. Line 96: It is unclear why the authors refer to anecdotal evidence?? Is there not empirically-based data? A reference to anecdotal data does not belong in the abstract

We agree with this comment and therefore have removed the sentence from the abstract. A WHO 2018 evidence review that suggests, based on key-informant data, that guidelines, training and support is needed to encourage appropriate use and we have mentioned this in the discussion.

15. Line 123: LICs should be spelled out first (i.e. “low-income countries (LICs)”).

Thank you; this has been changed in line with your recommendation.

16. Line 136 should be changed to just LIC.

Thank you; this has been changed in line with your recommendation.

17. Line 127: Missing a “)” after (10)

Thank you; this has been changed in line with your recommendation.

18. Figure 1: If possible, please rotate the top image so it is easier to read.

Thank you; this has been changed in line with your recommendation.

19. Lines 193-197: The method by which variables were chosen to be included in regression analyses needs to be made clearer. Why were these specific few variables chosen? For example, what does distance from hospital have to do with completion of cards? What were the “appropriate exploratory models”??

The regression analysis has been re-run. Characteristics were chosen based on clinical rationale, supported by both literature and the findings from the qualitative arm of the study. If variables were found to be too closely related following examination of cross-tabulations and the correlation matrices,resulting in multi-collinearity then they were excluded and clinical rationale was used to choose the most appropriate variable to include. This process has been explained in the methods section.

The methods section now contains a clearer description of the statistical methods and “appropriate exploratory models” has been removed.

20. Table 2 should have asterisk, like Table 1, indicating which variables were included in the modeling.

Thank you; this has been changed in line with your recommendation.

21. Line 247: Shouldn’t this be 79/250 not 251?

Thank you; this has been changed in line with your recommendation.

22. Line 334-335: Please expand on the studies referenced here. Did they have similar findings to the present study? Why yes or no?

The discussion has been significantly restructured to include a more detailed literature review and explicit comparison of the results to the literature.

23. Lines 354-367: What about the other significant sociodemographic factors? What hypothesis do the authors have for the association between sociodemographics (being a housewife, living close to the hospital, literacy) and card completion?

The results of the new regression analysis (where variables were selected based on clinical rationale from the literature and the qualitative arm of the study) have been explored far more extensively in the discussion and we tried to link them to the qualitative findings and the literature to explain the potential reasons for them.

24. Lines 361-362: This appears to be the only reference to the qualitative data in the discussion. The authors should expand on how their findings are related to previous research and/or describe more the relation between their quant and qual data

We have re-written a lot of the discussion section with heavy emphasis now placed on reconciling the quantitative and qualitative findings to build a picture. We have especially made use of the rerun of the logistic regression analysis to relate our findings to previous literature.

We have also rerun our literature review, which helpfully highlighted studies we had not previously commented upon.

25. Line 362: I believe this should be changed to “in our sample, 100%..”. “In fact” seems to indicate that staff always review 100%, which cannot be inferred from this data.

Thank you; this point has been removed from the discussion as it was considered to not be clear. Greater discussion of the qualitative findings has replaced it.

26. Line 381: Please expand on in what ways you would anticipate the situation to be different in rural Gambia and why.

Following previous discussions we have had with colleagues in The Gambia alongside added evidence provided by key-informants in the most recent WHO review of evidence on maternity cards, we have tried to expand on how the situation may be different in rural Gambia in the discussion of the limitations of the study. We make particular parallels with a study in Pakistan.

Since urban hospitals were sampled, the results may be less generalisable to rural Gambia than the urban Banjul area, although all rural areas around Banjul did refer women to these three hospitals. In Pakistan it was shown that maternity cards were more effective in rural than urban centres(40). On discussion with public health colleagues in the Gambia and reflecting on our study results that show women further from the hospitals had less complete documents, we would expect document use and completeness might be lower. This may be due to differing resource levels at the rural hospitals and clinics that might place more time-pressure on clinicians.

27. Line 387: What type of sub-group analyses were under-powered? What analysis of high-risk women was underpowered? Almost half of the sample was high-risk.

Thank you; the sentence in question was in reference to analysis performed in previous drafts that was not included in the final publication. Therefore it has been removed.

28. Line 392: Mother should be changed to mothers.

Thank you; this has been changed in line with your recommendation.

29. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

We have altered the formatting of the document in line with the information in the links you have provided. We have separated the supplementary materials and figures from the file, and labeled them appropriately.

Submitted filename: Responses to Reviewers.docx

Click here for additional data file.

18 Dec 2019

PONE-D-19-17140R1

The content and completeness of women-held maternity documents before admission for labour; a mixed methods study in Banjul, The Gambia

PLOS ONE

Dear Senior Clinical Lecturer Manaseki-Holland,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Both reviewers raised some remaining textual issues, and some minor points regarding the interpretation of your findings.

We would appreciate receiving your revised manuscript by Feb 01 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Astrid M. Kamperman

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #3: All comments have been addressed

Reviewer #4: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #3: Yes

Reviewer #4: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #3: Yes

Reviewer #4: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #3: Yes

Reviewer #4: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #3: This revised manuscript is a marked improvement from the original submission. It is much clearer and is synthesized very well. I only have a few outstanding concerns to be addressed:

1) I couldn't find S1 appendix and perhaps this information is already included there, but could the authors include the number of participants and the cadre of participants in the FGDs as they did for the SSIs in the body of the text?

2) I have a few comments on table 1. The number of children includes 6 twice- "5-6" and "6 or more." Also, the number of previous contacts includes 3 twice- "1 to 3" and "3 or more." Additionally, I don't fully understand the difference between the "1d" and "2d" notations. Can the authors clarify the difference between these? Also, what does "2d2d" in the high-risk section mean?

3) In the results section, lines 343-347, it would be useful to add odds ratios and specify what variables were significant in unadjusted and adjusted models.

4) In table 3, I'm not sure why/how the reference categories have p values. For example, for a categorical variable with 3 categories or levels, the p values would be looking at statistical significance comparing the 1st and 2nd categories/levels and the 1st and 3rd categories/levels, so there should only be p values for the 2nd and 3rd categories.

Reviewer #4: I commend the authors for a thorough response to my comments and revisions to the paper that greatly improve its readability and impact. The regression and qualitative methods and results are much more clearly explained and presented, and mixed-methods results have been well integrated. Below, I detail some minor remaining issues in the current revision.

Title – I believe you’d want to use a colon not semi-colon.

Line 113-116 – This seems more fitting in the discussion.

Qualitative methods: Traditionally you would want more than one coder of transcripts. The authors should explain why they only had one person code all transcripts and what was agreement on the 4 transcripts coded by other researcher?

Table 1:

• Ages seem to not include 20 (under 20 and 21+)

• Something is wrong with the education row. The numbers don’t line up to categories and the last category is Islamic…

• What is 2d 2d in missing high-risk?

Line 269 - It is not clear what documents were considered in this calculation. The authors state they were “not including loose scan or test result sheets” but then seem to consider these as “less frequently presented documents (ultrasound reports, prescription notes, scraps of 272 paper, child health reports, miscellaneous lab requests/results)”. What is a loose scan vs an ultrasounds, for example?

Line 282 – Why couldn’t the content of the card be assessed?

Line 301 – Why couldn’t the content of the referral be assessed?

Figure 4 does not make sense to me. I don’t understand the y axis, did women have hundreds of documents? This does not seem the best way to illustrate the point made in the text, that only the maternity card had met all 9 criteria.

Line 344 – Why don’t the authors mention the significant association of numbers of contacts throughout pregnancy in the results and discussion intro paragraph?

Table 3. I am unsure of what the p-value next to the reference category would mean. What is this a comparison of?

Table 4. [B] and [F] unnecessary considering the bolded row headings of barrier and facilitator.

Line 452 - It seems that the qualitative data DOES support this (quote from M6 in org barriers). The authors might want to reflect on how this was an issue identified here, so perhaps training should emphasize how helpful clinicians find the cards as a motivator. Or, perhaps a different incentive to complete them is needed.

Line 510 – ALMOST universally.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #3: No

Reviewer #4: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

1 Feb 2020

1. I couldn't find S1 appendix and perhaps this information is already included there, but could the authors include the number of participants and the cadre of participants in the FGDs as they did for the SSIs in the body of the text?

The S1 appendix provides some information about each of the hospital study sites, including the total number of nurses, midwives and doctors working at each site. We will make sure that this appendix is uploaded successfully on our next submission.

To address your point we have tried to clarify the wording of this sentence/paragraph to make it clearer what cadre of participants were in each of the FDGs.

2. I have a few comments on table 1. The number of children includes 6 twice- "5-6" and "6 or more." Also, the number of previous contacts includes 3 twice- "1 to 3" and "3 or more." Additionally, I don't fully understand the difference between the "1d" and "2d" notations. Can the authors clarify the difference between these? Also, what does "2d2d" in the high-risk section mean?

Apologies, these were errors when reconfiguring the table from the SPSS data – it was meant to say “more than 6” and “more than 3”, thank you for spotting it.

Apologies, the “2d2d” was another formatting mistake and has been corrected. Hopefully this now makes the notation easier to understand.

The 1d and 2d shows where missing values were in each subgroup – i.e. in high-risk, two women from hospital 2 had a their risk status missing from their questionnaire sheet, therefore in the total column two women had this data point missing. In ‘number of previous contacts’, one participant from hospital 1 and one participant from hospital 2 had this piece of data missing from their questionnaire sheets, which totals two women missing this data (the far right column).

Due to these mistakes we went back over the table to ensure all data was correct and have made a few appropriate changes. We have also changed Supplementary Material 1 due to the mislabelling of the hospital numbers.

3. In the results section, lines 343-347, it would be useful to add odds ratios and specify what variables were significant in unadjusted and adjusted models.

Thank you – we have added these to the paragraph (please see below).

In logistic regression analysis, being literate in English (OR 2.04 [95% C.I. 1.08-3.85]), having 1-4 children compared to having fewer or more (OR 4.4 [95% C.I. 1.04-18.07]), having more than 3 contacts with healthcare during pregnancy (OR 2.16 [95% C.I. 1.15-4.03]) were all positively and significantly associated with minimum criteria fulfilment. Travelling further than 1 hour to get to hospital (OR 0.34 [95% C.I. 0.15-0.74]) and attending a hospital other than the tertiary referral centre (Hospital 2 OR 0.45 [95% C.I. 0.19-1.02], Hospital 2 OR 0.22 [95%C.I. 0.08-0.60]) were negatively associated with minimum criteria fulfilment [Table 3].

4. In table 3, I'm not sure why/how the reference categories have p values. For example, for a categorical variable with 3 categories or levels, the p values would be looking at statistical significance comparing the 1st and 2nd categories/levels and the 1st and 3rd categories/levels, so there should only be p values for the 2nd and 3rd categories.

The p value against the reference category is the p value for the variable’s overall contribution to the model (as are the p values for binary variables). As the p values against other categories do not add information above that provided by the 95% Cis, they are redundant and have been removed from the table (but could be reinstated if the editor wishes).

5. Title – I believe you’d want to use a colon not semi-colon.

Thank you, we have now changed this in line with your recommendation.

6. Qualitative methods: Traditionally you would want more than one coder of transcripts. The authors should ex

plain why they only had one person code all transcripts and what was agreement on the 4 transcripts coded by other researcher?

It is considered good practice in qualitative research for more than one research to be involved in coding transcripts but there is no absolute requirement for all transcripts to be coded by another researcher (see for example N Gale et al BMC Res Methods 2013 where it is suggested that another researcher code four transcripts). The concept of interrater agreement is contested in qualitative research (C Pope BMJ 2000), and in this study the other coder contributed to the development of coding, improving the quality of the analysis. Time and resource precluded the double coding of all transcripts and a comment to that effect could be added if the editor wishes.

7. Table 1:

• Ages seem to not include 20 (under 20 and 21+)

• Something is wrong with the education row. The numbers don’t line up to categories and the last category is Islamic…

• What is 2d 2d in missing high-risk?

- We have changed this to ’20 and under’, thank you for spotting this mistake.

- The lining up of Table 1 when opened on a windows computer was flawed. We have now edited the table on a windows computer to make sure that the numbers line up. ‘Islamic’ category refers to home or small group based education about the Quran, and the student does not learn to read or write English which is the written language or commerce, business and general use. However as the person has exposure to some education we felt it is important to keep this category separate. We have inserted a foot note to say: renamed this as informal/formal Quranic education without learning to read English the national language

- This was a typo and has been correcting, thank you for spotting it.

8. Line 269 - It is not clear what documents were considered in this calculation. The authors state they were “not including loose scan or test result sheets” but then seem to consider these as “less frequently presented documents (ultrasound reports, prescription notes, scraps of 272 paper, child health reports, miscellaneous lab requests/results)”. What is a loose scan vs an ultrasounds, for example?

Thank you very much for making us aware of this statement that we agree lacks clarity. We recognised that the calculation also features in our abstract, so we returned to our raw data to ensure that our statement was accurate. The phrase ‘loose sheets’ remained there after an earlier draft and needed review.

We decided that to ensure clarity, we have calculated the value as ‘antenatal cards or referral sheets’ and now explain this clearly both in the paragraph and the abstract. We have made this decision to not include ‘less frequently presented documents’ (which were often ‘loose’) for two reasons. Firstly, these were the only two documents to include significant clinical information (the less frequently presented documents were therefore considered less clinically important to comment on when considering if a woman had presented with any documentation). Secondly, there was a data collection error where 40 participants did not have the presence ultrasound/test requests recorded. This was alluded to on line 323 (“35 women who had attended scanning were not able to be assessed for ultrasound scan presence in their documentation” – 5/40 did not attend scanning). We have added explanation of this to line 324 to make it clear that it was researcher error.

With regards to the paragraph you have commented on, one participant presented with no documents, but we are missing data as to whether she brought an ultrasound sheet with her or not. Because of this we cannot comment on the number of women who brought ‘no documentation’ of any kind. If the editors feel further clarity is needed regarding this matter, please let us know.

9. Line 282 – Why couldn’t the content of the card be assessed?

The content of the cards could not be assessed because although the mother could confirm that she had brought an antenatal card with her to hospital, the cards themselves were elsewhere in the hospital e.g. on the Neonatal unit with the baby.

We have added this to the text.

10. Line 301 – Why couldn’t the content of the referral be assessed?

The content on some occasions could not be assessed because some of the referral sheets that women recalled bringing with them were elsewhere on the ward rather than with the woman/unable to found by researchers.

We have added this explanation to the text.

11. Figure 4 does not make sense to me. I don’t understand the y axis, did women have hundreds of documents? This does not seem the best way to illustrate the point made in the text, that only the maternity card had met all 9 criteria.

We have removed the numbered labels on the Y-axis to make the bars represent the percentage of that type of document that met the minimum criteria (either 9 out 9, or 8 out of 9). We have also altered the X-axis title to simplify the message.

The minimum criteria fulfilment of ‘referral sheets’ and ‘other documents’ has been added to emphasise that only the maternity card met all 9 criteria.

We are happy to adapt or remove this figure if the editor wishes.

12. Line 344 – Why don’t the authors mention the significant association of numbers of contacts throughout pregnancy in the results and discussion intro paragraph?

Thank you, this has now been added (with odds ratios) to the appropriate paragraphs.

13. Table 3. I am unsure of what the p-value next to the reference category would mean. What is this a comparison of?

Please see response to comment on Table 3 above.

14. Table 4. [B] and [F] unnecessary considering the bolded row headings of barrier and facilitator.

Thank you, we have changed this in line with your suggestion.

15. Line 452 - It seems that the qualitative data DOES support this (quote from M6 in org barriers). The authors might want to reflect on how this was an issue identified here, so perhaps training should emphasize how helpful clinicians find the cards as a motivator. Or, perhaps a different incentive to complete them is needed.

This was an important point to have been brought to our attention, thank you. We reflected on this and have added appropriate comments on this in our discussion.

16. Line 510 – ALMOST universally.

Thank you, we’ve made this change.

Submitted filename: Responses to reviewers.docx

Click here for additional data file.

5 Feb 2020

PONE-D-19-17140R2

The content and completeness of women-held maternity documents before admission for labour: a mixed methods study in Banjul, The Gambia

PLOS ONE

Dear Senior Clinical Lecturer Manaseki-Holland,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses a final point.

Please remove Figure 4 from the text. You might add the figure it to the supplementary material.

We would appreciate receiving your revised manuscript by Mar 21 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Astrid M. Kamperman

Academic Editor

PLOS ONE

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

20 Feb 2020

Please remove Figure 4 from the text. You might add the figure it to the supplementary material.

Thank you for your comment regarding Figure 4. We have removed figure 4 and added it to the supplementary materials.

Submitted filename: Responses to reviewers.docx

Click here for additional data file.

21 Feb 2020

The content and completeness of women-held maternity documents before admission for labour: a mixed methods study in Banjul, The Gambia

PONE-D-19-17140R3

Dear Dr. Manaseki-Holland,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Astrid M. Kamperman

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

28 Feb 2020

PONE-D-19-17140R3

The content and completeness of women-held maternity documents before admission for labour: a mixed methods study in Banjul, The Gambia

Dear Dr. Manaseki-Holland:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Astrid M. Kamperman

Academic Editor

PLOS ONE