Use of pregnancy personalised follow-up in case of maternal social vulnerability to reduce prematurity and neonatal morbidity

Since the 80 s, social deprivation has been demonstrated to be a major contributor of inequalities in health [1]. It is now well established that socially deprived women experience more pregnancy complications with an increased risk small for gestational age newborn [2], stillbirth [2, 3], premature birth [4‐9], extremely premature birth (both spontaneous and induced [10, 11]). Therefore, neonatal morbidity and mortality are more frequent in socially deprived women [12, 13]. Several works aiming at explaining these discrepancies in term of adverse pregnancy outcomes reported that socially deprived women do not use properly prenatal care [14, 15]. Indeed, vulnerable women seems to address their pregnancy follow-up late, with a lower number of pregnancy consultations and often don’t realize the appropriate number of screening ultrasound [16‐18]. Therefore, it seems relevant to detect social deprivation in order to improve perinatal care use and reduce the impact of social vulnerability on pregnancy outcomes. Yet, studies aiming at evaluating the impact of measures to improve the perinatal care use for deprived women are rare. Three studies from the US have demonstrated that the detection of social vulnerability and an appropriate follow-up regarding this latter allowed to reduce stress factors such as interpersonal violence, addiction during pregnancy and improved perinatal care use in socially deprived women [19]. Moreover, two of these studies demonstrated that the companionship of these patients reduced the rate of SGA new-born [20] and premature birth in some sub-groups [21]. Consequently, the finding of new interventions to address social vulnerability is a major health issue. The aim of this study is to evaluate whether a personalized pregnancy follow-up (PPFU) to ease the access to both prenatal care and relevant interventions regarding the patient’s vulnerability would improve neonatal morbidity in case of maternal social vulnerability.

Between January 2020 and December 2021, 3958 women with social vulnerability who delivered a singleton after 14 gestational weeks were included in the study (Fig. 1): 686 in the PPFU group and 3272 in the standard care one. Patients in the PFFU group were younger and presented lower BMI (Table 1). Excessive and insufficient GWG were equally distributed between the two study groups. The PPFU group included more patients with a Sub-Saharan Africa origin. High medical risk level before pregnancy was more frequent in the PPFU group. Women included in the PPFU group presented a higher SDI with a larger number of patients presenting a SDI ≥ 3 (45% versus 19% in the standard care group, p < 0.001). Recent immigration, interpersonal violence during pregnancy and addictions both before and during pregnancy were more frequent in the PFFU group. Minor patients and patients with handicap were all included in the PPFU group. Linguistic barrier was more frequent in the standard care group. Patients with unwanted pregnancy were more prevalent in the PPFU group.

Inadequate PCU was more frequent in the PPFU group (Table 1). Women included in the PPFU group benefitted from more social worker intervention, addiction therapy, psychologic and psychiatric follow-up compared to the standard care group. The number of pregnancy follow-up appointments and hospitalizations was higher in the PPFU group whereas emergency consultations were equally distributed between the two groups.

Pregnancy complications were similar between the two groups (Table 2) and each complication taken separately was equally distributed (Additional Table 1).

Induction of labor was similar between the two groups (Table 2). Women included in the PPFU group delivered more often by cesarean section before labor whereas the emergency cesarean section rates were similar between the two groups.

Regarding neonatal outcomes, the premature birth < 37 GW rate were similar between the two groups (10.9% in the standard care group versus 8.6% in the PPFU group, p = 0.084) whereas the PPFU group unveiled a lower rate of premature birth < 34 GW (6.4% for the standard care group and 4.2% in the PPFU one, p = 0.038). SGA status was equally distributed between the group groups. Women in the PPFU group had a lower rate of NICU admission (6.7% for the PPFU group versus 9.7% for the standard care group).

After adjustment on the social deprivation index, maternal age, parity, BMI, maternal origin, high medical and high obstetrical risk level before pregnancy (Table 3), PPFU was associated with a reduced risk of premature birth < 37 GW (aOR = 0.66, 95%CI[0.49–0.89]), a reduced risk of premature birth < 34 GW, (aOR = 0.53, 95%[0.34–0.79]). There was no association between PPFU and SGA (aOR = 1.06, 95%CI [0.86 – 1.30]). Propensity score matching adjusted (psa) OR performed on the same variables unveiled similar results for premature birth < 37 GW (psaOR = 0.63, 95%CI[0.46–0.86]), for premature birth < 34 GW (psaOR = 0.52, 95%CI[0.34–0.78] and for SGA status (psaOR = 1.07, 95%CI [0.86 – 1.33]).

To our knowledge, this retrospective study is the first to evaluate a personalized patient-based intervention to address all types of social vulnerabilities with the aim of reducing poor neonatal outcomes. This work highly suggests that PPFU reduces the impact of social stress factors on prematurity and neonatal morbidity by improving medical care access and coordination for socially deprived women.

Indeed, patients included in the PPFU group had a significantly better access to any of the different care providers such as social care workers, psychologists, and addiction therapist. Even if inadequate PCU was more frequent at patient inclusion in the PPFU group, PPFU patients had more pregnancy follow-up appointments and were more often hospitalized (Table 1). This point suggests that PPFU patients had a more appropriate pregnancy follow-up given that the rate of pregnancy complications was similar between the two groups and might explain the differences observed in terms of prematurity. Moreover, maternal psychiatric disorders [27], pregnancy related anxiety in socially deprived women [28] and maternal stress [29] are established risk factors of premature birth. Therefore, the benefit regarding prematurity in the PPFU group can be explained by the difference in terms of psychologic follow-up, psychiatric follow-up, and addiction therapy.

The novelty of our personalized clinical approach comes from the facts that PPFU addresses all types of maternal social vulnerability and proposes customized pregnancy follow-up. Former studies aiming at reducing the impact of social vulnerabilities on neonatal outcomes in the US demonstrated that reducing psychosocial stress factor along with dietary counselling [20] or improving prenatal care access [21] reduces the rate of low birth weight and prematurity in the sub-group of Hispanic women [21]. Yet, within these studies, included patients were proposed similar pregnancy follow-up and only particular types of social vulnerabilities were addressed.

This work suggests that PPFU reduces prematurity regardless of maternal ethnicity. Yet, PPFU had no impact on SGA rates compared to the two studies carried out in the US [20, 21]. This point can be explained by the absence of dietary counselling within the PPFU. Indeed, inadequate GWG was similar between the PPFU and standard care group.

It is of importance to highlight that this study was carried out in high risk patients with high figures of social vulnerability, which was also the case for the two aforementioned studies [20, 21]. Indeed, a recent randomized control trial in the Netherlands testing an intervention aiming at detecting and addressing social deprivation in a low risk population that didn’t reach the expected figures of social deprivation [30], didn’t show any benefit regarding low birth weight and premature birth.

It is of note that the results are consistent with previous studies addressing social vulnerability and poor pregnancy outcomes (Table 3). Indeed, the increase of the social deprivation index was independently associated with premature birth (< 37 GW) (aOR = 1.16[1.04–1.30] for one point increase of the SDI), premature birth (< 34GW) (aOR = 1.28[1.11–1.49]) [2, 4, 5, 7, 16, 23]. In addition, these data are consistent with the impact of PCU on pregnancy outcomes unveiled in former studies [15, 16]. Indeed, the main differences between the two study groups resides in the access to both medical and non-medical prenatal care (Table 1). This suggests that the differences observed in terms of poor neonatal outcomes comes from the improvement of prenatal care use.

The main strength of this study is the combined accurate and thorough access to both precise data on the social vulnerabilities and medical factors thanks to the computerized patient folder. This was the result of the training of all the professionals of the maternity unit performed by the PPFU unit midwifes since 2017, in the detection of social vulnerabilities. Moreover, the sample size of socially deprived women is substantial compared to former studies conducted in France [10, 15, 31]. The access to medical and obstetrical history of the patients allowed to properly adjust on medical and non-medical factors in the multiple logistic regression and propensity score matching (Table 3).

Yet, this study is not without limitations. The retrospective and unicentric design might limit the generalization of the results. Indeed, the population studied in this work presented a high level of social deprivation with 54.2% of the included patients presenting a SDI ≥ 2 and 23.7% with a SDI ≥ 3. Moreover, the neighborhood of the maternity unit at study present a high level of social deprivation with a high rate of neonatal mortality [32]. Finally, the vulnerable women population included in this study presented a high medical risk level and a high rate of pregnancy complications (Table 1, and Table 2). Therefore, the application of the same protocol in a less socially deprived population with a lower medical risk might alter the impact of PPFU. Lastly, the retrospective design of this study lessens the strength of the results. Yet, testing this intervention in a prospective randomized manner would be ethically questionable given the impact of social vulnerabilities on pregnancy outcomes.

This work suggests that PPFU to address social vulnerabilities reduces prematurity and neonatal morbidity. These results could be explained by an improvement of relevant prenatal care access.

Further research should focus on characterizing the relevant non-medical needs according to social vulnerability profiles to improve both mother’s and neonates’ wellbeing.