Maternal multimorbidity during pregnancy and after childbirth in women in low- and middle-income countries: a systematic literature review

Background For every maternal death, 20 to 30 women are estimated to have morbidities related to pregnancy or childbirth. Much of this burden of disease is in women in low- and middle-income countries. Maternal multimorbidity can include physical, psychological and social ill-health. Limited data exist about the associations between these morbidities. In order to address all health needs that women may have when attending for maternity care, it is important to be able to identify all types of morbidities and understand how each morbidity influences other aspects of women’s health and wellbeing during pregnancy and after childbirth. Methods We systematically reviewed published literature in English, describing measurement of two or more types of maternal morbidity and/or associations between morbidities during pregnancy or after childbirth for women in low- and middle-income countries. CINAHL plus, Global Health, Medline and Web of Science databases were searched from 2007 to 2018. Outcomes were descriptions, occurrence of all maternal morbidities and associations between these morbidities. Narrative analysis was conducted. Results Included were 38 papers reporting about 36 studies (71,229 women; 60,911 during pregnancy and 10,318 after childbirth in 17 countries). Most studies (26/36) were cross-sectional surveys. Self-reported physical ill-health was documented in 26 studies, but no standardised data collection tools were used. In total, physical morbidities were included in 28 studies, psychological morbidities in 32 studies and social morbidities in 27 studies with three studies assessing associations between all three types of morbidity and 30 studies assessing associations between two types of morbidity. In four studies, clinical examination and/or basic laboratory investigations were also conducted. Associations between physical and psychological morbidities were reported in four studies and between psychological and social morbidities in six. Domestic violence increased risks of physical ill-health in two studies. Conclusions There is a lack of standardised, comprehensive and routine measurements and tools to assess the burden of maternal multimorbidity in women during pregnancy and after childbirth. Emerging data suggest significant associations between the different types of morbidity. Systematic review registration number PROSPERO CRD42018079526.


Background
Maternal multimorbidities affect millions of women during pregnancy and after childbirth and the burden of illhealth is expected to be highest in women in low-and middle-income countries (LMIC) [1]. For every maternal death, 20 to 30 women have morbidities related to pregnancy or childbirth [2,3]. More recent studies using new and comprehensive assessment tools suggest that the magnitude of maternal multimorbidity is much larger than previously estimated [4][5][6]. International targets and the Sustainable Development Goals have a new focus; in addition to preventing maternal mortality, improving health and well-being, as well as "survive and thrive" are the new goals [7]. There is international agreement that all women have the right to the highest attainable standard of health and well-being, also during pregnancy and after childbirth [7,8]. Estimates of morbidity have until now largely focused on acute and/or severe complications such as haemorrhage, sepsis and eclampsia [9]. The current definition of health is "a state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity" [10]. There are arguments that this definition needs to be re-formulated to consider health in a context of functionality, capacity, adaptability and the ability to perform activities of daily living despite having an illness or disability; but with a continued emphasis on the importance of the three domains of health: physical, psychological and social [11]. There is also debate that current definitions, measurements and timeframes for "multimorbidity", "co-morbidity", "morbidity burden" and related constructs are not well conceptualized [11,12].
Regarding maternal morbidity, a suggested definition is "any health condition attributed to and/or aggravated by pregnancy and childbirth that has a negative impact on women's well-being" [13]. In order to address all health needs that women may have when attending for maternity care, it is important to be able to identify all types of morbidities and understand how each morbidity influences other aspects of women's health and wellbeing during pregnancy and after childbirth. To date, lack of data exist regarding measurement and burden of disease described as "maternal morbidity", "maternal multimorbidity", or "maternal co-morbidity"; these terms are often used interchangeably; and there is uncertainty regarding the timeframe over which maternal morbidity impacts a woman's health and wellbeing. Additionally, there is limited understanding of best practices to measure different components of maternal ill-health and descriptions of morbidities, and if and how different types of morbidities are interlinked and associated.

Objective
A systematic review of the literature was conducted for studies from LMIC that measured two or more different types of maternal morbidity and/or associations with and between morbidities.

Methods
We included studies which assessed two or more types of maternal morbidity in women during and/or after pregnancy. For the purposes of this study we categorised maternal multimorbidity as physical (such as but not limited to medical, infectious, obstetric), psychological (such as but not limited to depression, suicidal ideation) and social co-morbidities (such as but not limited to domestic violence, substance misuse) [5]. We assessed tools that were used to collect data, including self-reported subjective measures; and/or objective measures such as clinical examination; and/or use of investigations for different types of maternal multimorbidity as reported by authors. We described how and what different types of maternal multimorbidity (physical, psychological, social) were measured and if there were any reported associations between these.

Data sources and search strategy
This protocol is registered in PROSPERO (CRD42018079526). Relevant articles published between January 2007 and December 2018 were identified using a structured search strategy in four electronic databases: CINAHL Plus, Global Health, Medline, and Web of Science. A search strategy was developed using thesaurus (including MeSH) and free-text terms for "maternal morbidity" and associated keywords, were used as main search terms. For each aspect of maternal morbidity ("physical", "psychological" and "social") search terms and related keywords were selected (Supplementary Table 1). Reference lists and bibliographies of key topic articles were also searched and any additional papers that met the inclusion criteria were obtained.

Inclusion and exclusion criteria
The study population was limited to women during pregnancy, childbirth or up to 12 weeks postnatal. Studies were excluded if: (i) they reported one type of maternal morbidity only, (ii) examined trend, risk factors or associations only without estimates of prevalence of types of morbidity, or (iii) reported severe or lifethreatening complications of pregnancy or childbirth that would require emergency obstetric care. The review was limited to studies from LMIC as defined by the World Bank. Language was limited to English.

Selection and data extraction
One researcher screened all titles and abstracts (MMc). A sub-sample (20%) was double screened by the second researcher (SZ). Evaluation of full-text papers was done independently by these two researchers with reasons for exclusion recorded and any discrepancies were discussed with a third researcher (NvdB). Information was extracted into a pre-designed summary table and included data on location of study, study dates, study design, study population, types of maternal morbidity, methods of measurement, timing (pregnancy phase) of the assessment and whether or not associations were reported (Supplementary Table 2). Throughout the review and extraction process, articles where uncertainty existed were discussed by all researchers to reach consensus.

Quality assessment
Appraisal of the quality of studies was conducted based on descriptions of maternal morbidities, sampling methods and completeness of data. Quality of evidence for each study was assessed using the Grading of Recommendations, Assessment Development and Evaluation (GRADE) tool adapted from the Critical Analysis Skills Programme (CASP) tool [14].

Data synthesis
A narrative synthesis approach was used to describe outcomes including: types of maternal morbidity categorised as physical (such as medical, infectious, obstetric), psychological (such as depression, suicidal ideation) and social (such as domestic violence, substance misuse); approaches used to collect data (self-reported or determined by healthcare providers); data collection tools used (standardised validated tool, or study specific); measurements of maternal morbidities; and reported associations (if any) between different types of maternal morbidities.

Results
By combining the search terms, 2840 studies were identified from the four databases and after screening for relevance, 58 were retrieved for full text review (Fig. 1). Upon applying the eligibility criteria, 38 articles met the inclusion criteria. Two studies were conducted by the same group of authors [15][16][17][18]. In these publications, the same methodology was reported in two papers, but there was a different emphasis on the results and outcomes reported per publication. For the purposes of this review, the first publication is referenced in the methodology section [15,17]. Both publications were included in the summary tables and measurements and/or associations for each publication are described in the results section. Most studies (92%; 33/36) were of medium quality, and the rest low quality (8%, 3/36).

Characteristics of studies
The 36 studies were from 17 different countries, with 15 from sub-Saharan Africa. Eleven were conducted in lowincome countries and six in middle-income countries (four lower-middle and two upper-middle income countries).

Physical morbidity
Twenty-nine studies reported on different types of physical morbidity; three of which assessed pre-selected populations including women with HIV [35,42] or women with gestational diabetes [38]. A variety of data collection tools were used, but generally not well described.

Associations between different types of morbidity
For physical morbidity, there was an association between increased psychological morbidity in women with obstetric complications (haemorrhage, infections, incontinence, prolonged labour, caesarean birth, low birthweight, stillbirth, neonatal death) ( Table 2) [6,26,41,45,50]. Women with gestational diabetes were not more likely to have psychological morbidity (depression) [38], but women with HIV were more likely to have social morbidity (domestic violence) [18]. Psychological morbidity was more common in younger women [40] and among women with social morbidities such as domestic violence [25,35], unwanted pregnancy [19,41,50] and poor social support (Table 2) [41]. For social morbidity, there was an association between women with substance abuse (alcohol) and domestic violence [48] ; and domestic violence was also associated with neonatal death [48] and maternal complications (Table 2) [46]. Due to heterogeneity, meta-analysis of associations was not possible.

Main findings
There is emerging evidence of a high burden of multimorbidity in women living in LMIC during pregnancy and after childbirth, as well as emerging evidence of associations between physical, psychological and social morbidities, suggesting that maternal morbidities are inter-linked. There is, however, still limited data about the strengths and direction of the associations between the different types of morbidities.
There was an apparent lack of standardisation of definitions and data collection tools used to measure maternal multimorbidities. The EPDS was the most common validated data collection tool to assess psychological morbidity in the studies, but with different cut-off scores to determine the risk of "depression" (ranging from 4 to 13) making comparisons difficult. Similarly, a variety of different validated data collection tools were used to assess domestic violence and/or substance abuse as components of social morbidity. Physical, psychological and social morbidities were often described as aggregates or summative measures, limiting comparability of findings.

Strengths and limitations
To the best of our knowledge, this is the first systematic review to assess maternal multimorbidities and types and levels of association between the different types. Many studies relied on recall of experience of morbidity and many primary data were symptom-based rather than "diagnosed". Only four studies triangulated self-reported symptoms with findings from clinical examination and/ or basic laboratory investigations. Assessments of measurements of ill-health based on self-reporting may be valid regarding ill-health as experienced by women, but do not provide accurate burden of disease estimates. No study described or used internationally recognised disease classifications to assess physical morbidity. Internationally recognised data collection tools were used to assess psychological and social morbidity, but these often used different cut-off scores making comparisons difficult. A limitation of this review is that studies that explored maternal multimorbidity using qualitative methodology were excluded.

Interpretation
Valid comparable measurements of maternal multimorbidity are limited to date, and this study confirms the need for a new approach and focus [70][71][72]. It will be important for future healthcare practice and research to agree and apply: (a) common identification criteria for maternal multimorbidity taking into account the different types of physical, psychological and social morbidity; (b) standardised and validated data collection tools that can be used in different languages and at all levels of healthcare; with, In models adjusted for sociodemographic factors and co-morbidities, all postpartum illnesses were associated with an increased relative risk of depressive symptoms in women by 6 months postpartum. These morbidities included uterine prolapse (RR 1.20, 95% CI 1.04-1. 39 Multivariate logistic regression showed that for pregnant women in Malawi, after controlling for parity and pregnancy stage, antepartum bleeding increased the odds of psychological morbidity 5-fold (OR: 5.01; 95% CI 1.60, 15.70; p = 0.006). Infective morbidity (i.e. for each additional infective morbidity) showed more than 2.5fold increase in the odds of having psychological morbidity (OR: 2.58; 95% CI 1.92, 3.47; p = 0.000). For Pakistan, there was a 56% increase in odds of psychological morbidity due to increasing burden of infective morbidity (OR: 1.56; 95% CI 1.36, 1.79; p = 0.000), and 78% increased odds due to increasing burden of non-infective morbidity (OR: 1.78; 95% CI 1.51, 2.11; p = 0.000), when controlling for the effect of complications during previous pregnancy. Complications during previous pregnancy, infective morbidity (p < 0.001), intra or postpartum haemorrhage (p < 0.02) were associated with psychological morbidity in both settings.

Psychological morbidity
Faisal-Cury 2009 [16] Obstetric complications were independently associated with common mental disorders in pregnant women.
Hanlon 2009 [45] Significant associations exist between pregnant women who report intimate partner violence and preterm labour, need for caesarean section, antenatal hospitalization and vaginal bleeding.
Natasha 2015 [38] There was no association between women with depression and gestational diabetes mellitus or other obstetric factors. However, pregnant women's level of literacy, poor household economy, poor relationship with husbands, and partner violence showed strong associations with depression and anxiety.
Prost 2012 [50] Unwanted pregnancy, small perceived infant size and stillbirth or neonatal deaths were all independently associated with increased risk of psychological distress in postnatal women. Loss of infants or unwanted pregnancies increased the risk of distress considerably (aORs: 7.06 95% CI: 5.51-9.04 and 1.49, 95% CI: 1.12-1.97).

Rees 2016 [47]
For pregnant women with any mental distress, adjusted odds ratios for four or more traumatic events and severe psychological abuse was 3.60 (95% CI 2.08-6.23); for four or more traumatic events and physical abuse 7.03 (95% CI 3.23-15.29); and for four or more traumatic events and severe psychological and physical abuse the adjusted OR was 10.45 (95% CI 6.06-18.01). For pregnant women who reported four or more traumatic events, and either physical abuse alone or in combination with severe psychological abuse, there was a 10-fold increase in depressive and other mental health symptoms.
Ukacukw 2009 [28] After multivariable adjustment, intimate partner violence intensity had a strong and statistically significant association with depression symptom severity for pregnant women.

Waqas 2015 [41]
Results of unadjusted log-binomial regression showed that unwanted pregnancy, prenatal depression and social support were associated with low birth weight.
Wong 2017 [42] Inferential analysis revealed that higher HADS scores were significantly associated with lower social support scores, rural background, history of harassment, abortion, caesarean birth and unplanned pregnancies (P < .05). Romero-Gutiérrez 2011 [46] Maternal complications were higher in pregnant women who experienced violence (30.2% vs 23.6%, p = 0.004). Pregnant women who experienced sexual violence had more maternal complications (43.2%), and pregnant women who experienced psychological violence had more neonatal complications (54.2%).

Stöckl 2010 [48]
Women's odds of drinking alcohol during pregnancy were significantly increased if they had experienced violence during pregnancy. Violence during pregnancy was also associated with having had a child or infant that died.
(c) validation of self-reported measurements of maternal morbidity compared to clinical assessment, investigations and diagnosis determined by healthcare providers [6,[70][71][72]. More recognition must be given that maternal morbidity is a complex concept with important associations between different morbidities. This has implications for screening and management of all different types of illhealth during pregnancy and after childbirth. There is a need to incorporate women's understanding, perceptions and lived experience of maternal multimorbidity into public health approaches to improve maternal health and wellbeing during pregnancy and after childbirth in LMIC [73,74].

Conclusion
To date a range of methods and tools have been used to assess maternal multimorbidity. Maternal multimorbidity estimates using these methodologies and tools, while useful as a guide, cannot be considered truly representative of the burden and range of maternal multimorbidity that have negative impact on women's wellbeing during pregnancy and after childbirth. The suggested WHO definition of maternal morbidity provides such a framework in principle, but challenges remain to map out comprehensive, feasible and acceptable assessment tools, approaches and timeframes [11]. Comprehensive and routine measurements of maternal multimorbidity are necessary to inform policy and program decisions and for resource allocation for antenatal and postnatal care [5]. Improved standardised measurements of maternal multimorbidity will also allow for comparison of the burden of disease across settings within and between countries. There is a need for a sustainable way to provide good baseline maternity care for all and targeted individualised care for women who need extra care to prevent development and progression of maternal multimorbidity.