General Practitioners’ Perceived Indicators of Vulnerability in Pregnancy- A Qualitative Study

Objective To explore general practitioners’ (GPs’) perceived indicators of vulnerability among pregnant women in primary care. Design A qualitative study with semi-structured in-depth focus group interviews. Setting General practices located in a mixture of urban, semi-urban and rural practices throughout the Region of Southern Denmark Main outcome measures Through qualitative analysis and systematic text condensation of the interview data, the following themes emerged: (1) obvious indicators of vulnerability - i.e. somatic or psychological illnesses, or complex social problems and 2) intangible indicators of vulnerability – i.e. identication depended on the GPs’ gut-feeling. From the GPs’ perspective, the concept of vulnerability in pregnancy were perceived as the net result of risk factors and available individual and social resources, with a psychosocial etiology was the dominant framework. Conclusions The GPs demonstrated a broad variety of perceived indicators of vulnerability in pregnancy; most importantly, the GPs were aware of a group of pregnant women with intangible vulnerability. Despite not tting into the GPs perceived concept of vulnerability, the GPs had a gut feeling that these women might be vulnerable. Misjudging the resources of pregnant women due to their physical appearance could delay the GPs’ identication of vulnerability. Future studies should explore the challenges GPs experiences when assessing vulnerability among pregnant women.


Introduction
Vulnerability among fertile women is increasing due to low psychosocial resources, such as social problems and mental health problems (1). Undetected vulnerability during pregnancy may result in increased risk of complications either during the pregnancy, during the birth, or throughout childhood (2). Vulnerability in pregnancy is described in the literature as psychosocial problems (i.e. history of anxiety or depression before or during pregnancy) or social problems (i.e. young age, lack of social support, being single, unemployed, low education level (3)(4)(5)(6)(7)(8), history of adverse childhood experiences, poor socioeconomic status (9), stressful life events during pregnancy (6), or a history of domestic violence or abuse (10)). Additionally, many vulnerable women of fertile age, report alcohol consumption above the high risk level of seven units of alcohol per week (1,11).
In antenatal care, vulnerability constitutes a major factor in the development of inequalities in maternal and perinatal health (12,13) and increases the risk of a debut or relapse of depression during pregnancy or a postpartum depression (3,5,8). Additionally, vulnerability in pregnancy is signi cantly associated with negative birth outcomes, such as; preterm birth, low birth weight, low APGAR scores (14) and adverse outcomes in childhood i.e. disturbed mother-child relation with risk of child neglect, emotional problems and symptoms of attention de cit hyperactivity disorder (15)(16)(17). Increased support in early pregnancy is crucial for vulnerable pregnant women which will help them bene t from the antenatal care system, and subsequently decrease related negative birth outcomes (18). Furthermore, interventions for decreasing perinatal mental health problems have been shown to be cost-effective (19). Therefore, vulnerable pregnant women are in need of extra support; during pregnancy, at birth and during the postpartum period (2). However, very few vulnerable pregnant women proactively seek help (20) and are therefore reliant on the health professionals' ability to determine vulnerability.
Over the past decades in Danish antenatal care all pregnant women are offered an early pregnancy consultation with their general practitioner (GP) between gestational week 6-10, with the purpose of evaluating the need for extra support during pregnancy by assessing the pregnant women's comorbid risks and psychosocial resources (2). However, a report from the Danish National Board of Health indicated that it was challenging for the GPs to identify vulnerability among pregnant women, since only 25% of the most vulnerable pregnant women seen at the specialized social obstetric units were referred by their GP (21). This nding is in line with challenges reported for GPs in the UK (22). In order to identify a vulnerable pregnant woman, the GP must rst understand which indicators implies vulnerability in pregnancy. Studies from the UK and Ireland have assessed GPs' perceptions of perinatal mental health problems (23)(24)(25), however the GPs' understanding of what indicates vulnerability in pregnancy was not explored.
This article is part of a project exploring GPs' perceived challenges in managing the care of vulnerable pregnant women. The aim of this study is to explore GPs' perceptions of what indicates vulnerability in pregnancy.

Design
The study was designed as a qualitative study and was conducted using semi-structured in-depth focus group interviews with GPs. The interview guide was developed by the research group, which consisted of four GPs (LBV, DEJ, RE, JS) and a questionnaire expert (LBP). Inspiration was gained from eld observations conducted by LBV in social obstetric units which handled vulnerable pregnant women. The Consolidated criteria for Reporting of Qualitative research (COREQ) checklist was used to ensure transparency (26).

Setting
In Denmark, general practices are organized as either single-handed practices (one GP per practice) or partnership practices varying in size (two-ten GPs per practice). The health care system in Denmark is tax funded and free of charge for the patient. Most Danes (99%) are registered with a GP of their own choosing, and the GP functions as a gatekeeper to secondary care (27).
Three antenatal care visits and one postnatal visit are offered by GPs. Danish antenatal care is a wellstructured collaboration which usually involves GPs, midwives, health visitors (who are specialized municipal nurses) and the obstetric units. If the GP perceive a pregnant woman as vulnerable, the specialized social-obstetric units and municipal social workers can be involved depending on the severity of vulnerability (2). During the rst pregnancy consultation, the GP completes the mandatory prestructured national pregnancy health record concerning; lifestyle habits (smoking, alcohol intake and use of addictive drugs), socioeconomic status, previous obstetric history and known somatic or psychological disorders. The national pregnancy health record helps the GP assess somatic and psychosocial vulnerability, which if detected, will require extra support during pregnancy. Like most services provided in general practice, the consultation is free of charge to patients. However, patients of ethnic origin who are not uent in Danish and have lived in Denmark for longer than three years, are charged a fee for translator assistance (28).

Data collection Selection and recruitment of participants
The study aimed to recruit a purposive sample of participants representative of the GP population with respect to; gender, years of experience, practice type and various practice localizations throughout the Region of Southern Denmark. Respondents were recruited via letter, telephone, e-mail and snowball sampling. However due to slow recruitment, the end sample consisted of a convenience sample of 20 GPs representing only partnership practices. Participant demographic details are shown in Table 1.  (5) 6-10 years (2) 11-15 years (5) > 15 years (5) Single-handed practices (0) Partnership practices (20) Urban area (5) Semi-urban area (11) Rural area (4) Female (12) Male (

Data management and analysis
All interviews were audio recorded and transcribed verbatim by LBV and uploaded to data processing software (NVivo) for coding and data organization. Open inductive coding with systematic text condensation (29) was used to ensure an in-depth investigation of themes and subthemes, where themes could freely emerge from data without being imposed by prior theory. Systematic text condensation is a method for thematic cross-case analysis and consists of the following steps: 1) gaining overview and elicit preliminary themes, 2) developing code groups from preliminary themes, followed by identi cation and sorting of meaning bearing units, 3) establishing subgroups which exemplify important aspects of every code group by condensing the contents and 4) synthesize the condensed text from each code group (29). After the themes were discussed among the authors, LBV conducted the initial coding. The following stepwise analysis was conducted by LBV in correspondence with RE. The research team discussed and re ected on the ndings until consensus was reached.

Research Ethics
The study was carried out in accordance with the European data protection rules with informed consent from participants obtained before initiation of the study. The Research & Innovation Organisation at University of Southern Denmark approved the data processing activities regarding this project (journal number 10.307). According to Danish regulations, the study did not need approval from a health research ethics committee, since no research on human tissue or other biological material was performed.

Availability of data and materials
The dataset generated and analyzed during the current study are not publicly available due to them containing information that could compromise research participant consent but are available from the corresponding author on reasonable request. Data supporting the ndings of this study was used under a license granted speci cally for the current study and therefore is not publicly available according to the data protection regulations of Danish Data Protection Agency

Results
In general, the GPs perceived vulnerability as the net result of both risk factors and available individual and social resources. A psychosocial etiology appeared to be the dominant framework held by the GPs when conceptualizing vulnerability in pregnancy, and the availability of a social network was perceived as being important. As illustrated by one male GP when he said: "Vulnerability has to be understood in a social context and not only in an individual context" (male GP > 45 years old) As shown in Fig. 1, the GPs' expressed a variety of conditions indicating vulnerability in pregnancy, originating from different comorbid conditions; somatic disease, psychiatric disease and social problems. Even though the GPs were not asked to classify indicators of vulnerability in pregnancy in levels, the GPs' perceived speci c indicators as being more obvious and classical; whereas, other indicators were perceived as being intangible. Therefore, during the data analysis process of the GPs perceived indicators of vulnerability, to main themes emerged; 1) obvious indicators of vulnerability and 2) intangible indicators of vulnerability.

The obvious indicators of vulnerability in pregnant women
The GP perceived obvious indicators of vulnerability in pregnancy could be organized into three categories; 1) social problems, 2) psychiatric diseases and 3) somatic diseases.
Obvious indicators of vulnerability related to social problems were; signs of deprivation with known social cases in the system -i.e. known history of being neglected in childhood or history of having children forcibly removed, known low intellectual-or mental resources, low level of education and often being unemployed. Being a single pregnant woman was perceived as a possible indicator of vulnerability; especially if the woman was very young, with a broken relationship to the father or plans of parenting alone, and simultaneously having poor social resources. As one female GP said: "They're young, haven't known their partner for very long, have no education, no plans for their future and are often unemployed. We may know the family already, as a low social class family with low intellectual resources." (female GP, > 45 years) Poorly integrated women with an ethnic background were also perceived as vulnerable; since these women usually had a poor understanding of the language, the culture and local health system procedures. Lack of translator assistance made it di cult for the GPs to evaluate these women's resources and guide them through the medical system. The GPs perceived that it was often necessary to refer them to the social-obstetric care units but experienced that the women's poor economic and structural resources prevented them from attending the care units.
Other obvious indicators of vulnerability were related to known mental health problems, such as minor psychiatric disorders -i.e. anxiety, depression, attention de cit disorder or personality disorder(s)) with or without prescribed antidepressants, or known history of abuse of alcohol, drugs or addictive medicine. Whether the above pertained to the pregnant woman or her partner, they were perceived to indicate vulnerability.
A history of chronic somatic comorbidity or severe obstetric complications were perceived an indicator of vulnerability due to the risk of complicating pregnancy. Additionally, presence of chronic somatic disease was perceived to increase the degree of vulnerability, since they naturally caused a higher level of stress from worries.
Finally, the addition of several indicators of vulnerability as psychiatric diseases concomitant with coping problems from disabilities or chronic diseases were perceived to increase the degree of vulnerability.
"I had a patient with a hearing disability who was pregnant(...), however her real challenge is her many psychiatric challenges as she had been mentally unstable with poor self-care and di culties managing social challenges(..) plus, I don't think her intelligence level is very high" (female GP,< 45 years) Conversely, some GPs reported being positively surprised by patients perceived as being obvious vulnerable. This was the case in situations where pregnant women, typically the young women or women with psychiatric disease(s), appeared to grow with the task and became brilliant mothers. However, mostly it demanded great social support from cross-sectoral collaborators in both the social-obstetric and social care system in the community.
"I would not have believed that this girl with mild schizophrenia would succeed in getting her daughter home from the hospital and now I observe normal mother-child interaction when she visits my clinic" (male GP > 45 years)

The intangible indicators of vulnerability
The GPs reported cases of women whom they perceived having indicators of an intangible degree of vulnerability, which could not be related to any prior known somatic or psychiatric diseases or obvious social problems. These indicators of intangible vulnerability perceived by GPs could be abnormal behavior or odd contact of the pregnant woman, which triggered the GPs' gut feelings' that something is wrong, and that the pregnant woman might have an undisclosed psychiatric disease or a deviant personality.
The GPs were guided by their gut feeling in cases where no prior doctor-patient relation existed, due to lack of continuity or the patient was new in the clinic. As a female GP said: "I had this girl, who was a new patient and came for removal of her anticonception implant which we removed. Something was odd in the contact with her, and it triggered my attention -something was wrong. After thoroughly reading through her le, I discovered that her child was forcibly removed from her home by the social authorities a few years ago" (female GP, > 45 years) The GPs perceived several indicators of intangible vulnerability in pregnancy related to minor social challenges -such as coping problems with being pregnant, relationship challenges when having a baby, low selfcare, low perceived intelligence, low threshold of stress, perceptions of being lonely without support from spouse. However, there was a gradual transition to normal challenges of motherhood and parenthood. As none of these indicators of intangible vulnerability are visual in the patient les, these women could go undetected until developing obvious signs of depression. As a female GP said.
"I had a pregnant woman where everything in her patient le looked ne, but when she came for the ve weeks examination of her child I discovered that there was no eye contact with the baby, and subsequently I realized that the mother had a severe postpartum depression -and I thought 'why didn't I discover that?'" (female GP > 45 years) Interestingly, the GPs described being guided by the woman's appearance in their evaluation of the patient's resources. If a woman appeared with normal interpersonal behavior and well dressed, they were less likely to elaborate further on the woman's resources, which could delay the identi cation of vulnerability. This was often the case among the higher educated couples.
"They were this sharp looking couple driving an Audi and carrying designer sunglasses -completely streamlined upper class people you know. It was late in her pregnancy when she rst caught my attention, as it appeared how horrible she felt, and that they simply could not embrace the changes that were awaiting them" (female GP < 45 years) Especially, the GPs' were misguided by a woman's physical appearance when they had no pre-existing relation to the patient.
"They can easily trick you. I remember a woman, very nicely dressed, coming for her rst antenatal consultations. She was a new patient and I had no previous medical record on her. We went through her pregnancy record nice and easy, I asked about alcohol use and she said there was none. At the end of the visit she asked how should go about getting her alcohol treatment transferred to the local alcohol rehab center. She was not currently using alcohol, and therefore had answered no to all the questions. It's hard when we don't know them. Based on her appearance I had no idea that she was in alcohol rehab" (female GP, < 45 years) The GPs often acknowledged women with intangible vulnerability retrospectively, as they presented with a higher frequency of child consultations for minor things, such as simple colds, as they were insecure in judging the child's needs.

Statement of principal ndings
The GPs perceived vulnerability in pregnancy as the net result of psychosocial stressors and availability of individual personal and social resources, and classi ed vulnerability according to its obviousness. Especially, severe somatic comorbidity, known complex social problems, mental health problems and history of abuse were obvious to the GPs as vulnerability indicators. However, some pregnant women were perceived as being vulnerable due to intangible signs of vulnerability which evoked the GPs' gut feeling, and they felt unsure of what course of action to take. This was often the case when the women seemed to have normal resources, and there were no pre-existing doctor-patient relations.
Strengths and weaknesses of the study Strengths of this study were the semi-structured qualitative approach using focus group interviews and using COREQ criteria to ensure transparency (26). The study sought to reach a high information power(30) by continuing with interviews until reaching a study sample large enough to answer the research questions. The use of convenience sampling due to slow recruitment, could be a weakness of this study. Approximately 60 GPs were contacted whereas we recruited twenty GPs. A frequent reason for decline was due to lack of time. Only GPs from partnership practices participated, and the GPs that participated might have been the ones with a natural interest in the topic, thereby limiting the external validity. Though it was not an inclusion criterion, the recruited respondents varied broadly in their experience with conducting pregnancy consultations. This indicates transferability to GPs regardless their interest in the eld of antenatal care. Moreover,, we achieved diversity regarding seniority, gender and practice location from urban, semi-urban and rural areas. Therefore, the external validity was assessed as being fairly satisfactory. The

Findings in relation to other studies
This study adds knowledge to a sparsely covered area, about GPs' perceived indicators of vulnerability in pregnancy. Previous studies focused on GPs' perceptions of perinatal mental health problems, where vulnerability was classi ed as a risk factor of perinatal mental health problems. However, GPs' understandings of indicators of vulnerability were not explored (23-25, 31, 32).
Studies have shown that GPs are more aware of postpartum depression, and that intrapartum depression is underdiagnosed by GPs, with up to 50% missed cases (24,25,31). Studies comparing health care professionals found that health visitors were more knowledgeable and aware of perinatal depression than GPs and midwifes (31), and that midwifes lacked the necessary knowledge, skills and con dence to provide mental health care to pregnant women (33). Contrarily, our study differed in that it focused on GPs' perceptions of vulnerability and revealed that GPs had a broad clinical perception of indicators of vulnerability.
This study's ndings on GPs having gut feelings indicating an intangible vulnerability in a pregnant woman is in line with ndings from other studies, which described gut feelings among GPs as a sense of alarm that somethings is wrong, even if no objective argument was present (34,35). The GPs used their gut feeling as a compass in uncertain situations and trusted this feeling to guide them in their decision making. It was outside of the scope of this study to examine the predictive value of GPs' gut feelings, but other studies have found a general positive predictive value based on GPs' gut feelings(36).
In conclusion, this study has demonstrated how GPs perceives indicators of vulnerability in pregnant women in levels according to their obviousness. Obvious indicators of vulnerability in pregnancy was previously known social problems, psychiatric diseases or chronic somatic diseases, where a psychosocial etiology was the most dominant framework. The GPs described intangible indicators evoking their gut feeling that a pregnant woman might be vulnerable, such as perceived low intelligence, low threshold of stress, coping problems of being pregnant and missing support. However, these intangible indicators could be missed in women appearing with normal resources, especially when the GP had no relation to the patient. If GPs were more aware on conditions or situations indicating intangible vulnerability among pregnant women, it might increase their diagnosis of possible intrapartum depression. Further evidence is needed on the challenges GPs experience when assessing and addressing vulnerability (both obvious and intangible vulnerability), and when managing the care for vulnerable pregnant women through engagement in cross-sectoral collaboration.
Changes might be needed in the organization and structure of antenatal care in general practice and in the cross-sectoral antenatal care collaboration to ensure that proper attention is paid to women with intangible vulnerability. Declarations