.Soc. Sci. Med. Vol. 36, No. 4, pp. 495-507, 1993 Printed in Great Britain. All rights reserved

0277-9536193 $6.00 +O.OO Copyright 0 1993 Pergamon Press Ltd

THE LATIN AMERICAN TRIAL OF PSYCHOSOCIAL SUPPORT DURING PREGNANCY: A SOCIAL

INTERVENTION EVALUATED THROUGH AN EXPERIMENTAL DESIGN

ANA LANCER,‘* CESAR VICTORA,~ MAGDA VICTORA,* FERNANDO BARRD~,* UBALDO FARNOT,’ JOSE BELIZAN~ and JOSE VILLAR~

‘Division of Research on Woman and Child’s Health, Centro de Investigaciones en Salud Publica, Instituto National de Salud Publica. Av. Universidad 655, Sta. Maria Ahuacatitlln, C.P. 62508 Cuernavaca, Morelos, Mexico, *Departamento de Medicina Social, Universidade Federal de Pelotas, Brazil, 3Matemidad America Arias, Ministerio de Salud Publica, Cuba and 4Centro Rosarino de Estudios

Perinatales, Argentina

Abstract-Perinatal health problems are a public health priority in Latin America. Among the identified risk factors, psychological and social conditions play a crucial role. However, care during pregnancy and delivery in the region is usually hospital-centered and does not address women’s psychological and social conditions. The preeminence of research on perinatal health, along with the necessity for testing interventions that represent alternative models to improve women’s health, gave the Latin American Network for Perinatal and Reproductive Research grounds to develop a multicenter randomized controlled trial to evaluate a program of social support and health education during pregnancy.

The conceptual framework for this study was based on an ecological model of social support, i.e. a model in which social support and health education play a synergistic role and are meant to modify stressful situations and negative health-related behaviors. The target population consisted of women attending obstetric hospitals before the 22nd gestational week, at high psychological and social risk (n = 2236).

The intervention consisted of four to six home visits, carried out by social workers, and had four main components: the reinforcement of pregnant women’s social support network, emotional support, health education, and the improvement of health services utilization. The main foci of the intervention were determined after an ethnographic study was carried out to identify stress-producing situations and needs for support during pregnancy. Besides the home visits, the program also offered a hot-line, an office in the hospital, a specially designed poster and booklet, and a ‘guided tour’ of the health institution. Since this was a multicenter trial, the program’s standardization was a crucial methodological aspect that was achieved through the training course for the home-visitors team. Biological and psychosocial outcomes were measured in both experimental and control groups at the 36th week of gestational age, post-partum and at the 40th day after delivery.

The attributes of the multicenter population showed an important variability, reflecting differences in the countries or hospitals’ population prevalent attributes. The results of the program’s implementation were analyzed, demonstrating that home visitors adapted topics discussed during the interviews to the women’s conditions and the stage of pregnancy during which the visit took place.

In this paper, we demonstrate the feasibility of measuring social interventions through an experimental design; we discuss the benefits of developing an explicit conceptual and analytical framework; we show the possibility of applying a social program to a numerous and heterogeneous population; and we propose a methodology for the development of a social program which is specific to a population’s needs, standardized and flexible at the same time.

Key words-social support, perinatal health, clinical trials, multicenter studies

INTRODUCTION

During recent years, perinatal events have begun to attract the attention of researchers and decision- makers in Latin America. This is mainly because of the proportional increase in perinatal morbidity and mortality as a consequence of the reduction of infant mortality caused by infectious diseases [l], and because of the absolute increase in the number of births in developing countries where, by the year

*To whom reprint requests should be addressed.

2000, approx. 85% of births and 95% of perinatal deaths will take place 121.

At present, perinatal care in Latin America has a substantial ‘medical-curative-institutional’ orien- tation. Indeed, the present model for meeting health needs is, in general terms, centered on hospitals and based on medical doctors, mostly oriented to surveil- lance of fetal well-being, and does not include enough health promotion programs and preventive measures. Thus, the health-care system gives only limited cover- age to the population’s necessities, and it is generally unable to adapt itself to individual needs. The idea

495

496 ANA LANGER et al.

of comprehensive care for the mother and child, including the mother’s social and psychological needs, has not been sufficiently recognized [3].

The preeminence of research on perinatal health, along with the necessity for testing interventions that represent alternative models to improve women’s health, gave the Latin American Network for Perinatal and Reproductive Research (LANPER) grounds for developing a social support program intended to answer the following question: whether or not a preventive and support program within the present medical system could improve pregnancy outcomes, maternal well-being and satisfaction?

In this paper we stress the need to measure the effects of social interventions applying an exper- imental design, i.e. a randomized controlled trial; we demonstrate the benefits of developing an explicit conceptual and analytical framework; we propose a methodology for the development of a social support program, specific to a population needs, standardized and, at the same time, flexible in its application; and, finally, we show the feasibility of applying a social support program to a large and heterogeneous population of four different countries.

THE EFFECTS OF SOCIAL SUPPORT NETWORKS ON PERINATAL HEALTH: LITERATURE REVIEW

The negative effects of poor living conditions, social disorganization, and rapid social changes on health have long been recognized. However, before the 1970s the concepts of social network or support did not appear in the epidemiologic literature [4]. In the early 1970s published results suggested that some special social situations protected people from recognized risks. In the perinatal health field, Nuckolls called the psychological personal resources ‘psychological assets’ [5] and Cassel wrote about

‘social support’ [6]. Nuckolls et al. [5] evaluated correlations between

‘life-events’ and obstetric complications, including low birthweight (LBW). They found that 33% of women with high stress and high psychological assets had pregnancy complications, compared to 91% in the group with high levels of stress and low psycho- logical assets. Norbeck and Tilden [7] assessed stress, social support and other psychological variables in healthy pregnant women, and related these to their performance during delivery and the health of the newborn. They reported a significant inverse association between social support and obstetric complications.

Pascoe et al. [8] found that the help pregnant women receive in performing their daily tasks is one of the best predictors of birthweight. Reeb et al. [9] and Ramsey et al. [lo] demonstrated that inadequate family functioning and stressful events during preg- nancy are important risks factors for LBW. Working with pregnant teenagers, Koniak and Griffin [ll] found that some of the strongest predictors of an

adequate mother-child attachment were the total support received during gestation and the extension of the social support network.

In conclusion, observational studies show an association between stress, anxiety and lack of support during pregnancy and negative perinatal outcomes. Effective support and some psychological characteristics of the mother seem to have a protec- tive effect.

Even if a relationship between social support and health seems to be demonstrated, the possibility of a confounded association can not be ruled out through observational studies. Indeed, the underlying psycho- logical and socioeconomic characteristics of the mother could be the condition associated with both support and health. The best means of testing for this possibility is by implementing experimental inter- vention studies in which persons are randomly assigned to different support conditions.

Elbourne et al. [12] performed a meta-analysis of the published and ongoing studies in which interven- tions related to social and/or psychological support were evaluated in terms of biological and psycho- social outcomes (Table 1). In general terms the interventions consisted of home visits to pregnant women by nurses, midwives or lay persons, during which some kind of emotional support was offered, as well as counseling and education [13-191. Only one study in this review included support during labor and delivery [20]. In one project [15] the intervention included as an additional component, some financial assistance. The enrolment criteria were

Table I.

Instrumental/operative delivery 0.91 0.7&-1.07 Preterm delivery rate 1.06 0.82-l .36 LBW 0.92 0.77-1.10

Source: Adapted from Elboume D. er al. Social and psychological support during pregnancy. In Effective Care in Pregnancy and Childbirth (Edited by Chalmers I., Enkin M. and Keirse M.). Oxford University Press, Oxford, 1989.

Psychosocial support during pregnancy 497

being primiparous [15,20] or to have had a previous low birthweight baby [14, 16,18, 191.

In relation to the psychological and behavioral outcomes, women enrolled in the social support programs were less likely than those in the control group to feel unhappy, nervous and worried during pregnancy, were more likely to have good communi- cation with medical staff, felt a greater sense of control during delivery and were more satisfied with antenatal care [ 19,21,22,23].

During delivery, women in the experimental group were more likely to be accompanied, more satisfied with intrapartum care and less worried. Mothers who had received support during pregnancy were more likely to breastfeed immediately after birth and during the following 6 weeks, felt unwell less often and were less worried about their newborns’ health [15, 16, 19,22,24].

Effects on biological outcomes were not conclusive (Table 1). Nevertheless, some of the studies demon- strated that women in the intervention groups were less likely to use pharmacological methods of pain relief, and had a reduced incidence of Cesarean section [16, 19,221. The meta-analysis did not show any effect on preterm delivery and LBW, though in some of the reported trials modest positive results were obtained [IS, 161.

A conceptual problem may be contributing to this failure to identify specific impacts on biological out- comes: the concept of ‘social support’ is not clearly established in the literature. Some general definitions have been suggested, such as “a moderator of life stress” [6] . “and/or its health consequences” [25], “the resources provided by other persons” [26]. Berkman and Breslow [27] give a more detailed definition, describing social support as the amount of

INTERVENTION

High risk It pregnant women

Program:

Increase of support network

Emotional support

Health education

Reinforcement of adequate

health services utilization

emotional and practical support a person receives from his/her ‘social networks’ (i.e. the kinds of social ties a person maintains with other individuals and the community at large). The lack of accuracy of these definitions is a serious obstacle to comparing studies in this field and to making a complete systematic review.

An additional important aspect to consider is that, with only one exception [28], all the trials were performed in developed countries, where women’s needs of social support may be less intense than in other settings. However, it is important to point out that some of the studies were focused on minority or disadvantaged groups [15, 191. It seems reasonable to think that an intervention such as social support might have a stronger effect on women with greater needs and under high stress conditions, thus making it easier to identify causal relationships. In this study we have attempted to overcome some of the main limitations which we identified in other studies in this field.

ANALYTIC AND CONCEPTUAL FRAMEWORK

For this study, we based our conceptual framework on the ecological model of social support, developed by Olds and Kitzman [29] (Fig. 1). According to these authors,

the ecological model attempts to explain how factors at different levels of proximity to the mother and child (e.g. behavioral, biologic, psychologic, sociologic, and economic) interact to form a system of influences on maternal functioning. On a practical level, the ecological model leads the home visitor to examine simultaneously maternal per- sonal resources, social support, and stresses in the home, family and community that can facilitate or interfere with optimal health-related behaviors during pregnancy.

PROCESS OUTCOMES

FINAL OUTCOMES

Social support I

, L-1 / \ /I--

1 // Psychological Health-

distress related

t

I behaviors l--_7

A

Knowledge about

pregnancy and delivery

Fig. 1. Analytic and conceptual framework.

498 ANA LANGER et al.

In this study, women at high risk of being under stress during pregnancy-as a result of previous perinatal problems, bad socioeconomic conditions, poor health-related habits and/or lack of an adequate social support network-were identified as the target population.

The definition of ‘social support’ that we adopted was Caplan’s [30], who describes it as “a supportive system of formal and informal relationships through which an individual receives the emotional, cognitive and material support necessary to master stressful experiences”. This comprehensive and explicit definition was adopted here in order to measure natural social support networks and to develop the intervention. Thus, the program included four components: reinforcement of the available support network, emotional support, health education and the improvement of health services utilization.

The mechanisms through which the intervention was expected to influence perinatal outcomes were the raising of social support, the decreasing of psychological distress and the increasing of women’s knowledge about health promotion and danger signals during pregnancy and delivery. These modih- cations would influence maternal and newborn’s con- ditions directly and/or through ameliorating mothers’ health-related behaviors. Positive behavioral changes (i.e. adequate health services utilization, smoking cessation, resting to reduce strenuous physical activity, nutrition improvement, etc.) would mainly affect biological outcomes, but would also have an effect on satisfaction with the reproductive process and perinatal care. The expected decrease in psycho- logical distress should have an important and direct effect on psychological outcomes and might also have some straight effect on biological results [26].

The rationale for gathering the strategies of social support and health education in one program is not new. Along with Israel and Shurman, we believe that “programs can be more effective if they focus on reducing stress in addition to teaching individuals how to cope with stressful experiences. Health edu- cation interventions can also be aimed at strengthen- ing the moderating factors of social support and control that may reduce stress and improve health” [3l]. Both ingredients play a synergistic role and reinforce the whole program; thus, it may be useless to try to identify the most important component [29]. Since the relationship between stress and LBW may be mediated (or confounded) by inadequate health- related behaviors [32], Olds and Kitzman [29] looked for some empirical evidence of the importance of adding a health educational component to social support interventions: they analyzed four random- ized clinical trials [15, 16, 18, 331 to disentangle which background characteristics and which changes in intervening factors make population more sensitive towards social support, and found that the inclusion of a health education component was crucial to achieve positive results.

Even though we are aware of the limited effects demonstrated by previous social support and health education interventions [ 12,341, we estimate that two elements in the LANPER project increase the possi- bihties of obtaining positive results: the high-risk conditions of the women recruited, which made them especially suited for a support program, and the combination of social support and health education in the same intervention.

Finally, the development of an explicit frame- work, and the measurement of the variables included in it, will probably help to clarify the mechanisms through which social interventions affect health outcomes.

THE INTERVENTION STUDY

The design adopted for this study was a randomized

controlled trial (Fig. 2). It was carried out in four research units located in the cities of Rosario (Argentina), Pelotas (Brazil), La Habana (Cuba) and Mexico City (Mexico).

Screening and population characteristics

Considering that high-risk women usually do not attend our antenatal clinics before the second trimester of pregnancy, patients were screened in the four participating centers between the 15th and the 22nd week of gestation. Establishing this cutpoint inevitably eliminated women that usually seek care later in pregnancy and that probably are at the highest risk; however, it was decided that for the intervention to be effective, a minimum of I5 weeks was needed.

Some strong negative biological conditions were considered as exclusion criteria: cardiovascular or renal diseases, hypertension, history of cerclage, Rh( -), mental disorders or other major chronic illnesses that might interfere with pregnancy and usually require special care. Selection was performed through the application of a screening questionnaire that included the following variables: previous low birthweight or preterm babies; previous fetal, neo- natal or infant deaths; age < 17 yr; body weight < 50 kg; height < 1 SO m; low family income (defined through local standards); < 3 yr of schooling; smok- ing and/or alcohol consumption; single, separated, divorced or widowed.

After positive screening for at least one of these risk factors, a baseline questionnaire was filled out including complete information about reproductive history, family structure and functioning, fears and expectations, and available social support. Further- more, the Spielberger Scale-that has been already validated in Spanish-speaking population-was applied to measure pregnant women’s anxiety [35]. This information was used as a baseline measure and as input to plan some specific aspects of each woman’s home-visit program, allowing for a more personalized intervention.

Psychosocial support during pregnancy 499

Evaluation - 36 weeks

Gestational weeks

_.........__...._.___...._______..._________..._________......____._................... 15 22 // 21-23 25-27 29-31 33-35 36 Birth 40 days

Fig. 2. Flowchart of the intervention study.

The number of women with positive inclusion criteria was 2236, with a refusal rate of < 1%. After informed consent was obtained, patients were ran- domized. Ethical problems were not identified, since women in the control group received the routine prenatal care, and the program was not expected to have any negative side effects.

The information on whether a woman belonged to the control or the intervention group was not known to any other member of the research group and/or the health institution, except for the health visitors who were in contact with the intervention group and the supervisor. Contamination risk among women in the sample was low, since all hospitals in which the intervention was implemented were crowded and opportunities for women to interact very infrequent.

Preparation of the intervention

The general plan of the intervention was outlined following the results of an ethnographic study which was implemented to find out about women’s needs, fears and expectations during pregnancy. The ethno- graphic study consisted of nonstructured interviews of women at different stages of pregnancy, carried out by psychologists participating in the study.

The ethnographic study input, together with the research team’s a priori ideas about what content the intervention should have, led us to the design of the home-visit program. The input of the ethnographic study guaranteed the specificity of the program to the particular population’s characteristics, and avoided a common and unfortunate feature in health research: the lack of correlation between researchers beliefs and preconceptions, and the target population’s needs.

Standardization of the intervention

Since this was a multicenter study, the intervention should fit each center’s population risk conditions and needs and at the same time be similar in the four centers, in order to be comparable. Thus, standardiz- ation was a crucial methodological aspect and it was achieved through the training course for the team of home visitors. The course lasted 4 weeks, 6 hr a day, and included theoretical and practical instructions, using role-play on how to conduct the visits and respond to possible personal or social situations.

For this training course, a manual was prepared in which topics to be discussed during each visit were described in great detail. The practical part consisted of a ‘decision-making flowchart’ by which all home visitors in the four centers would react in a similar way in the same situations. Consequently, the train- ing manual ensured that a consistent, standardized program would be applied, even if the prevalence of each specific condition were different in the participating countries.

The standardization procedure that was developed for this particular trial could be a useful model for other studies in which complex interventions must be implemented under highly-controlled conditions in heterogeneous populations.

Description of the program

The core of the intervention consisted of four planned home visits, at approximately the 22nd, 26th, 30th and 34th weeks of gestation (Fig. 2). The number of home visits could be increased up to six when needed. The visits were performed by specially trained social workers or nurses, as was the case in Cuba. They were all female, and the selection

500 ANA LANGER et al.

criteria included their interest, empathic attitude, and capacity to explain health-related issues.

In the home visits, the pregnant woman, the visitor and a so-called ‘support person’ participated. The support person was chosen by each woman at the time of the enrolment, to accompany her during rhe interviews and other activities that were part of the intervention. Each woman was completely free to choose her support person which was, in some cases. the husband/boyfriend. and in others the mother, a sister, friend or neighbor.

An active participation of the support person was promoted throughout the pregnancy in order to increase the emotional and material support received by the woman. Thus, during the home visits the support person was encouraged to understand the woman’s needs, participate in the decision-making process. help to solve specific problems, promote healthy behaviors and encourage prenatal care at- tendance. The duration of the visits varied between I and 2 hr.

The subjects to be discussed during each inter- view were adapted to the corresponding time of pregnancy. The home-visits plan was, nevertheless, extremely flexible. In each interview, the visitors developed specific activities related to the four com- ponents of the intervention (Fig. 3).

Reirzforcenzmt of social support network. This was the main objective of the intervention. The first part of the home visits consisted of encouraging the pregnant woman and her support person to talk about the evolution of pregnancy, their worries and their doubts. family structure, neighbors and friends. Concrete ways were identified of getting their collab- oration in order to fulfil the women’s needs. The opinions of the support person were considered care- fully and given much weight. The visitor suggested very practical types of assistance for the pregnant woman: helping her with heavy domestic duties, care of small children, home care at the time she would have to go to the hospital etc. However, in some cases, reinforcement of the support network was achieved by discussing family communication troubles and by improving relationships within the family.

When the women’s current natural support net- work was unsatisfactory, the establishment of links with other people (i.e. colleagues at work, neighbors, relatives) was suggested. In some cases, when attempts to reinforce or create a support network were unsuccessful, the visitor herself became the main support person.

Emotional support. The visitor’s empathy, her inter- est in listening to and discussing whatever the preg- nant woman and/or her support person wanted to talk about, her role as a link between the woman and the hospital, along with her advice, made the visitor herself a vehicle for transmitting emotional support. This component was particularly relevant when the ‘natural support network’ was very weak.

COMPONENTS ACTIVITIES

- Reinforcement of social support network

* 4-6 home visits including the patient, her “support person” and the home visitor

- Emotional support

Knowledge about pregnancy and delivery

- Reinforcement of adequate health services utilization

. Conversation about woman’s “natural social support” network and needs

l Involvement of social support network in concrete and helpful tasks

l Visitor’s empathy * Visitor’s interest l Visitor’s role as a

link with the hospital

l Specific topics l Anti-smoking and

anti-drinking programs * Poster l Booklet

l Information about hospital characteristics and rules

l Hot-line l Guided tour in the

hospital ’ Consultation service

Fig. 3. Intervention components and activities.

Knowledge about pregnancy and delivery. During the home interviews, the visitor talked to the preg- nant woman and support person about health related topics, such as the normal development of pregnancy and the risks related to smoking, drinking, strenuous physical activity, inadequate nutrition etc. The visitor explained pregnancy complications (bleeding, preterm labor, physical signs of toxemia etc.) and the normal signs of labor. Aside from the more general topics, the information discussed during the visits was intended to be very specific to each woman’s particular situation. For certain very specific prob- lems like smoking and drinking, specially designed interventions were implemented.

In order to reinforce social networks and health education, a poster was given during the first visit in which pregnancy was represented as a path. Positive health behaviors and the role of the support person were clearly described (Fig. 4). The poster was ac- companied by a small pamphlet in which information on pregnancy-related danger signals was given, as well as the project’s hot-line number, the name of the visitor and some useful supplementary information.

CU

IDA

A

T

U

BE

BE

D

ESD

E

AH

OR

A

V

Fig.

4.

Po

ster

gi

ven

to

wom

en

in

the

expe

rim

enta

l gr

oup

,_

__

d.-Y

I_m

-.,l-

--

. ,,,

._

_

_ ^

. . .

.- I

_

- _

,,

>

r” I

..Y

-

- -

.I

. _-

502 ANA LANCER et al.

In synthesis, these activities would contribute to positive changes in health-related behaviors, to timely recognition of problems that could be treated, and to greater understanding of how to respond to stress- ful circumstances. Additional expected benefits were the reinforcement of the women’s ability to ask her doctor for accurate information, and to enable the support persons to participate more in health-related decision-making.

Reinforcement of adequate health services utiliz- ation. Complex institutional organization, lack of in- formation and various psychological factors usually interfere with an adequate utilization of available health services. In this program, an attempt was made to reverse this condition by providing information about the characteristics and rules of the health-care institution. A telephone ‘hot-line’ and a consultation service were offered by which pregnant women could contact the project’s supervisor, and discuss their doubts with her.

To diminish the fear of unfamiliar situations, a ‘guided tour’ of the hospital was organized period- ically. Women and their support persons were invited, around the 36th week of gestation, to trace the path they would follow at the beginning of labor.

Finally, the program functioned as a link between women and the health-care institution. Through this, we tried to improve hospital facilities utilization and to empower women in their role as clients, encouraging them to be more active and demanding.

Evaluation and outcomes

Process outcomes were measured in the interven- tion and control groups at the 36th week of gestation. Final outcomes were evaluated post-partum and 40 days after delivery (Table 2). 36th week and 40th day data were obtained through questionnaires filled out by ‘blind’, specially trained interviewers. Post-partum information was obtained from clinical records. Attrition rate was calculated for each outcome measurement. Losses were almost identical in the control and experimental groups and general attrition

Table 2. Process and final outcomes

36 gestational week questionnaire @recess oufcomes) Social support Anxiety Knowledge about pregnancy and delivery Health-related behaviors

Pm-partum form (JTnal oulcOmeS ) Birthweight Gestational age Newborn’s condition Intrapartum interventions Length of labor

40th day post-partum quesrionnaire (final outcomes) Newborn’s condition Puerperal morbidity Breastfeeding Anxiety Social support Health services utilization Mother’s satisfaction

rates were lower than assumed in the sample size calculations.

Quality control of the intervention and evaluation visits

In order to test the reliability of the information collected through the evaluation questionnaires and the post-visit reports, and to ensure that the visits were carried out as planned, a random sample of 5% of the 36 weeks and 40 days post-partum question- naires were repeated by a ‘blind’ interviewer. Further- more, a random sample of 5% of the visits were observed by one of the fieldwork supervisors (without previous announcement) who completed a duplicate version of the post-visit report.

Qualitative study of the women’s perception of the program

The data collected through the evaluation ques- tionnaires did not include the women’s own views of the program and its effects. It was therefore decided that a parallel evaluation process, using qualitative methodology, would be carried out in each country. Thus, in each center a trained supervisor observed visits number 2 and 4 of fifteen women, and filled out an observation checklist for each visit. This checklist was compared with the plan for the visit and the post-visit report completed by the social worker. In order to get information about the women’s perception of the program, the same women were interviewed using a semi-structured instrument after the 40th day post-partum questionnaire was administered.

IMPLEMENTATION OF THE PROGRAM

Population’s risk factors

The criteria for trial entry was the presence of one or more of the main risk factors for LBW delivery. The distribution of these risk conditions by center and in the whole sample was obtained (Table 3). The mean number of risk factors was 2.1 for the whole population.

Women included in this study were not representa- tive of the population of pregnant women in the study areas, since they were selected from those attending a limited number of antenatal services at a given gestational time. However, some results probably indicated the prevalent local conditions.

The proportion of LBW in each center varied, being particularly low among the Cuban sample, as was the case for previous infant death and the rest of the biological indicators, except for low maternal weight. This may reflect the higher levels of health that prevail among the Cuban population. Previous fetal death was more frequent in the Mexican health- care institution, a third-level hospital to which high risk population has greater access.

Adolescents (age 17 or under) were more common within the Argentinean sample. Proportions among the other three centers’ populations were lower and

Psychosocial support during pregnancy

Table 3. Proportion of women according to risk condition, by center Center

503

Risk condition

Argentina (%)

N = 522

Previous LBW Previous fetal death Previous infant death Age <I7 Height < 150 cm Weight < 50 kg Low income Crowding (4 or more people per room) Schooling < 3 yr Lack of husband/boyfriend Smoking

21.1 8.0 7.2

25.7 12.2 20.9 -

41.0 9.6

19.5 21.3

(%) (%) N=486 N = 608

Mexico

(%) N=620 N = 2236

17.9 10.8 20.1 17.7 9.1 4.5 13.6 9.5

13.6 2.8 13.7 9.7 11.9 13.8 14.2 16.3 8.1 10.2 28.2 15.2

35.4 39.3 16.0 27.7 53.5 20.4 54.0 42.6 10.5 22.4 20. I 23.5 12.3 0.8 2.9 5.9 16.3 12.7 18.9 16.8 36.8 28.1 7.9 22.8

Total

Number of risk factors 2.8 2.1 1.6 2.1 2.1

very similar. Low maternal weight and height were and neighbors. The pregnant women’s mother, included in the screening questionnaires as specific father, siblings and other relatives accompanied risk factors for LBW and as proxies of low socio- them in a small proportion of the visits. This variabil- economic status. Nevertheless, in our sample these ity confirmed the convenience of allowing a free two variables were not associated with each other: election of the support person rather than limiting the greater proportion of maternal height of < 1 SO m this role to pre-defined people. The interest that was found in the Mexican hospital population and the pregnant women and support persons demon- women that weighed < 50 kg were more frequently strated during the home visits was very high found in Cuba and Brazil. (Table 4).

Income levels were lower in Brazil and Mexico. In Argentina, national economic instability made it im- possible to establish clear cutpoints. The proportion of women with <3yr of schooling was very low in Cuba and higher in Brazil.

Lack of partner was a risk factor existing in a similar proportion in the four centers, somewhat lower among the Cuban sample. Smoking had a very high prevalence among Cuban and Brazilian samples, apparently reflecting the frequency of this habit among the general population.

In summary, the distribution of the risk factors varied greatly within the whole sample. Heterogen- eity may be a potential benefit for the study since, if positive results were demonstrated, differences among the participating centers would greatly increase the external validity of the study. However, this variabil- ity could also hide some effects in the multicenter sample since the same program could produce different results on different populations. Stratified analyses by risk factor will be performed to avoid the effect of different risk factor distribution.

To trace the development of the program, specific items about the topics discussed were included in the post-visit forms which the social workers completed. The mean number of themes addressed during the interviews was about 50% of the total of possible topics. The visitors were not supposed to approach all topics, since the intervention was to be adapted to the particular circumstances of each case.

In the following tables, the percentages of visits in which each topic was discussed are presented. ‘Discussion’, in this context, meant to ask about, listen, inform, give advice or make suggestions. When percentages were very similar in all visits, only the mean was included in the tables.

Topics related to communication with the partner, relatives, friends, etc. were addressed in a high pro- portion of the visits (Table 5). Even though these were central issues, the figures did not reach 100% because when there was no perceived trouble in relation to communication, this subject was only mentioned.

The following will describe another essential aspect of the application of the program: how visits were developed in practice.

Discussion about home, family and work was crucial, in order to define concrete ways of giving support to the pregnant women. This point was included in a high proportion of all visits.

Characteristics of the home visits

Support persons were present in 66.4% of the visits (Table 4). Thus, in 35% of the interviews women were alone, nullifying the expected involvement of the support person. In these cases, the chances of getting a positive result may have diminished.

The women’s partner was present in 34.7% of those visits in which the support person participated. The second most common companions were friends

Table 4. Characteristics of the home visits, all centers

Presence of ‘support parson’ 66.4% Presence of women partne? 34.7% Pregnant women ‘very interested 93.7% ‘Support person’ very interested’ 86.6% Number of topics discussedb 17.8

(5.5)

‘Only for visits with presence of ‘support person’. bMean number of topics discussed out of a total of 34

possible topics (standard deviation in brackets).

504 ANA LANCER ef al

Table 5. Proportmn of wsits III whtch topics related to reinforcement of soctal

support network and emotional support were discussed, all centers

I. Communication 84.0%

2. Home, family and work related-topics 90.3%

3. Normal bio-psycho-social changes during 1st and 2nd trimester

1st visit

2nd visit

3rd Gsit

4th visit

4. Normal blo-psycho-social changes during 3rd trimester

1st visit

2nd visit

3rd vi71t

4th visit

5. lntegratmn of the newborn

1st “IS!,

2nd VI\II

3rd Vl\ll

4th visilt

6 ldentitication of problems durmg pregnancy

Ict vl\lt

2nd VISIT

3rd visit

4th visit

7. Stress reductmn during pregnancy

8. Previous delivery experiences

lat visit

2nd visit

3rd visit

4th visit

9. Expectations about d&very

1st et

2nd visit

3rd visit

4th visit

96.9%

64.3%

29.4%

15.6%

16.1%

58.5%

XI.X%

7 I I %

30.2%

46.2%

71 3%

X9.4%

94.6%

83.9%

80.8%

78.9%

83.0%

21.8%

17.6%

20. I %

66.4%

24.2%

29.1%

39.4%

96.7%

Issues related to bio -psycho -social changes during pregnancy were divided into those of the first and second trimester, and those of the last. These topics could be considered as a part of the second com- ponent of the intervention, i.e. knowlege about pregnancy and delivery. However, we are including them as a strengthening of social support activity because it was an essential strategy in reinforcing the women’s support network: discussion about changes during pregnancy was a way of improving the support persons’ understanding of the process the women were passing through.

The frequency of conversations about these changes was high when it coincided with the phase of pregnancy in which they took place; i.e. first and second trimester changes were approached in a high proportion of first visits. This can be considered as an evidence of the adaptation of the program contents to the different pregnancy stages: in other words, to the moment in which women and support persons were interested and needed to understand what was happening.

Conversation about the integration of the newborn into the family (i.e. psychological, social and oper- ational changes that might be expected after delivery) was considered a useful way of mitigating the event- ual disruptive effect of the presence of a new member in the family. This topic was included in about 90% of the last visits.

Identification of problems during pregnancy was particularly common in the first visit, during which the social worker had to identify the pregnant women’s main troubles. Even though the proportion

of meetings in which this topic was addressed dimin- ished after the first visit, the frequency remained very high, indicating the consistent updating of trouble- recognition the visitors provided.

The importance of and the strategies for stress reduction during pregnancy were approached in more than 80% of all visits (Table 5). Discussion about previous delivery experiences increased towards the end of pregnancy. A similar trend was seen in expec- tations about delivery and preparation for breast-

feeding, delivery and post-partum. Discussion about complications during pregnancy and delivery was in- cluded in a high percentage of visits, with the contents adapted to the stage of pregnancy in which interviews took place (Table 6). General health related behaviors (i.e. diet and weight, rest and activity) were discussed in almost all the visits. Smoking and drinking habits were approached only when needed, with a stronger emphasis during the first meetings. Discussion about sexual life occurred in about half of the visits. Consid- ering the cultural idiosyncrasy of Latin American populations, the frequency is significant. Adequate use of antenatal care, emergency rooms, institutional services and the project’s hot-line were discussed in a very high proportion of all visits (Table 6).

These results showed that conversation about health-related topics was more frequent than dis- cussion about social support, probably due to a ‘professional bias’ among the home visitors who felt much more at ease talking about concrete subjects than discussing aspects related to fears, anxiety, doubts, and non satisfied needs for support.

Psychosocial support during pregnancy 505

Table 6. Proportion of visits in which topics related to health education and health services utilization were discussed, all centers

I. Complications during 1st and 2nd trimester 1st visit 2nd visit 3rd visit 4th visit

2. Complications during 3rd trimester 1st visit 2nd visit 3rd visit 4th visit

3. Complications during delivery 1st visit 2nd visit 3rd visit 4th visit

4. Maternal diet and weight 5. Rest and activity 6. Smoking

1st visit 2nd visit 3rd visit 4th visit

7. Alcohol 1st visit 2nd visit 3rd visit 4th visit

8. Medical drum

98.9% 66.2% 31.4% 15.1%

19.2% 65.0% 89.9% 71.5%

26.3% 37.2% 46.9% 93.8% 95.6% 96.3%

45.8% 29.5% 23.7% 20.3%

39.8% 14.8% 12.2% 9.3%

53.1% 9. Hygiene d&g pregnancy

IO. Preparation for breastfeeding 1st visit 2nd visit 3rd visit 4th visit

1 I. Preparation for delivery Isi visit 15.1% 2nd visit 23.7% 3rd visit 33.6% 4th visit 96.4%

12. Preparation for post-partum experience 1st visit 2nd visit 3rd visit 4th visit

13. Sexual life 14. Incentive to antenatal attendance 15. Adeauate use of antenatal care 16. Emergencies and use of health services 17. Emergencies and use of the hot-line

2.4% 5.1%

13.4% 93.4% 48.6% 96.4% 97.0% 96.0% 95.5%

70.7%

28.4% 37.3% 90.8% 68.8%

In general terms, there was a high correlation between the program’s implementation and what was planned and taught during the training course. An instrument such as a pre-coded, post-visit record, with its correspondent analysis is mandatory for social interventions in which the personal character- istics of people involved can substantially change the program when applied under real conditions.

DISCUSSION AND CONCLUSIONS

The seriousness of perinatal problems, along with the lack of resources dedicated to health care in Latin America, makes it absolutely essential that new programs be carefully evaluated in terms of feasibility and effects on health, before being subject to con- sideration by decision-makers.

Randomized clinical trials are generally proposed as the most adequate design to evaluate medical interventions implemented under highly controlled

conditions [36]. This design is seldom used to measure the effects of social interventions. After reviewing 532 articles dealing with health education programs, Loevinsohn found that only 67 were program evalu- ations. Of these, only 45% used a control group and 21% were controlled studies with a sample size > 60 individuals [34]. This should be an important con- sideration for researchers, since the lack of a control group eliminates the possibility of measuring before and after changes, and exaggerates the perceived effectiveness of an intervention [34]. Through this study, we hope to encourage social scientists to use more rigid evaluation methodologies which are usually applied only in biomedical research.

In this trial, we gathered the strategies of social support and health education even though, according to the literature review, neither of these have demon- strated definite effects on perinatal health conditions [12,34]. However, the combination of both strategies may have produced a synergistic effect and increased the effectiveness of the intervention [29].

The adoption of an explicit conceptual and analyti- cal framework, along with the metisurement of the concomitant variables have made a significant advance towards elucidating the mechanisms through which social support influences biological and psychosocial perinatal outcomes. Simultaneously, the large sample size of the LANPER project will likely contribute to establishing the effects of social support on perinatal health. According to the literature re- view, in only three trials the population was of similar proportions to ours [18, 37,381. Large sample size is very important in overcoming the gap between qualitative and epidemiological studies. A large sample size is essential in obtaining comparable base- line characteristics in both groups and reducing the selection bias. The power of randomization is pro- portional to the sample size for reducing disparities between groups.

The characteristics of the women that participated in this study correspond to those of a population with a great need of support during pregnancy (Table 3). The application of the program to this type of target group could also contribute to the debate about the effects of social support. If results are not obtained from this large sample of women with important needs, this strategy would have to be seriously ques- tioned as a means of improving perinatal health. On the other hand, if positive outcomes were demon- strated in this study, the variability in the population enrolled in four Latin American countries would give the results a great external validity. Since effects might be stronger in some specific groups, results have to be looked for in special groups stratified by specific risk conditions and by participating center. Sample size and variability in this study allow for stratification.

There are various methodological contributions made by this investigation. The use of an ethno- graphic study as a starting point for the design of the

506 ANA LANCER et al.

intervention allowed the development of a program adapted to population needs. The standardization procedure was successful in maintaining comparabil- ity among centers without loosing the flexibility of the intervention. We recommend the standardizing methodology that was developed in this study for complex interventions, in which characteristics of the population and members of the research team may influence the program’s implementation. Quality control and post-intervention assessment by means of specially designed instruments and qualitative methods are indispensable in controlling potential variability among centers. These activities are highly encouraged for all kinds of program evaluations, but especially for multicenter studies.

From the start study was planned to be duplicated and institutionalized, provided that the results were favorable. Being different from other studies [15] in which interviews were very frequent and started at the early stages of gestation, we decided to employ a program of only four to six home visits started at the 20th week of the pregnancy. A program like this could be too weak to influence perinatal health. Nevertheless, a more intensive program could not be introduced into an institutional program given cost constraints.

The fact of having conducted this research in Latin America has some important implications: first, we designed and tested an intervention specifically ad- dressed to the conditions of our population, instead of ‘importing’ the conclusions of other studies in this field. Secondly, given the detailed and carefully developed methodology in this study, results should have a strong impact on whether or not to put resources into social support interventions. Thirdly, we had great success in carrying out a complex multicenter study. We hope that these results will stimulate the collaboration between researchers in the region. Lastly, we implemented a feasible preventive program which could be applied in ob- stetric hospitals without modifying current prenatal attention.

The exhaustive report given in this paper should permit study replication, criticism, and the inclusion of LANPER’s study results in reviews and meta- analyses in the field of social support and health education. Finally, the effort to apply a hard evalu- ation method to a social intervention should be seen as an attempt to overcome the traditional gap be- tween social and biomedical research. Along with Mosley and Chen [39], who pointed out the need of an interdisciplinary approach to the study of child survival, we consider comprehensive conceptual frameworks, designs and methodologies to be mandatory in order to propose effective responses to complex problems.

Acknowledgemenf-This project was supported by the International Development Research Center of Ottawa, Canada, through grant 3-P-87-0232-05.

I.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

REFERENCES

Langer A., Lozano R. and Bobadilla J. L. Effects of Mexico’s economic crisis on the health of women and children. In Social Responses to Mexico’s Economic Crisis ofthe 1980s (Edited by Gonzalez de la Rocha M. and Escobar A.), pp. 1955219. Center for U.S.-Mexican Studies, San Diego, Calif. 1990. Potts M., Janowitz B. and Fortney J. (Eds) Childbirth in Developing Countries. MTP Press, Boston, Mass. 1983. Oakley A. The Captured Womb: A History of the Medical Care of Pregnant Women. Blackwell, Oxford. 1984. Berkman L. The relationship of social networks and social support to morbidity and mortality. In Social Support and Health (Edited by Cohen S. and Syme L.). Academic Press, New York, 1985. Nuckolls K. B., Cassel J. and Kaplan B. Psychosocial factors, life crisis and the prognosis of pregnancy. Am. J. Epidemiol. 15, 431441, 1972. Cassel J. The contribution of the social environment to host resistance. Am. J. Epidemiof. 104, 107-123, 1976. Norbeck J. and Tilden P. V. Life stress, social support, and emotional disequilibrium in complications of preg- nancy: A prospective, multivariate study. J. Hlth .soc. Behav. 24, 3&46, 1983. Pascoe J., Chessare J., Bough E., Unch L. and Ialongo N. Help with prenatal household tasks and newborn birthweight: there is an association? J. Deve/op. Behat>. Pediafr. 8, 207-212. 1987. Rebb K., Graham A.. Zyzansky J. and Kitson G. Predicting low birthweight and complicated labor in urban black women: A biopsychosocial perspective. Sot. Sci. Med. 25, 1321Ll327, 1987. Ramsey C. N., Abel1 T. D. and Baker I. C. The relationship between family functioning, life events, family structure and the outcomes of pregnancy. J. .family Pratt. 22, X2-527. 1986. Koniak G. and Griffin L. The relationship between social support, self-esteem, and maternal-fetal attach- ment in adolescents. Res. Nurs. Hlfh 11, 269-278, 1988. Elbourne D., Oakley A. and Chalmers I. Social and psychological support during pregnancy. In E&live Care in Preanancv and Childbirth (Edited bv Chalmers I., Enkin M‘: and_Keirse M.). Oxford University Press, Oxford, 1989. Sokol R. J., Woolf R. B., Rossen M. G. and Weingarden K. Risk, antepartum care and outcome: Impact of a maternity and infant care project. Obsret. Gynec. 56, 15c-156, 1980. Heins H. C. and Nance N. W. A statewide randomized clinical trial to reduce the incidence of low birth- weight/very low birthweight infants in South Carolina. In Prevention of Preterm Birth (Edited by Papiernick E., Breart G. and Spira N.). INSERM, Paris, 1986. Olds D. L., Henderson C. R. Jr, Tatelbaum R. and Chamberlin R. Improving the delivery of prenatal care and outcomes of pregnancy: A randomized trial of nurse home visitation. Pediatrics 77, 16-28, 1986. Oakley A., Rajan L. and Grant A. M. Social support and pregnancy outcomes. Br. J. Obstet. Gynaec. 97, 155-162, 1990. Spencer B. and Morris J. The family worker project: Social support in pregnancy. In Prevention of Preterm Birth (Edited by Papiernik E., Breart G. and Spira N.). INSERM, Paris, 1986. Spencer B., Thomas H. and Morris J. A randomized controlled trial of the provision of a social support service during pregnancy: The South Manchester Fam- ily Worker Project. Br. J. Obster. Gynaec. 96, 281.-285, 1989.

19.

20.

21.

22.

23.

24.

25.

26.

21.

Psychosocial support during pregnancy 507

Dance J. A social intervention bv link-workers to 28. Klaus M. H., Kennel1 J. H., Robertson S. S. and Pakistani women and pregnancy ouicome, 1987. Cited in Elboume D., Oakley A. and Chalmers I. Social and psychological support during pregnancy. In Efictive Care in Pregnancy and Childbirth (Edited by Chalmers I., Enkin M. and Keirse M.). Oxford University Press, Oxford, 1989.

29.

Sosa R. Effects of social support during parturition on maternal and infant morbidity. Br. med. J. 2930, 585-587, 1986. Olds D. L. and Kitzman H. Can home visitation improve the health of women and children at environ- mental risk? Pediatrics 86, 108-l 16, 1990. Caplan G. Support Systems and Community Mental Health. Behavioral Publications, New York, 1974. Israel B. and Schurman S. Social support, control and the stress process. In Health Behavior and Health Edu- cation. Theory, Research and Practice (Edited by Glanz K., Marcus Lewis F. and Rimer B.), p. 188. Jossey-Bass, San Francisco, Calif., 1990. Newton R. W. and Hunt I. P. Psychosocial stress in pregnancy and its relation to low birthweight. Br. med. J. 285, 1191-1194, 1984. Dawson P., Van Doominck W. J. and Robinson J. L. Effects of home-based, informal social support on child health. Develop. Behav. Pediatr. 10, 6367, 1989. Loevinsohn B. Health education interventions in devel- oping countries: A methodological review of published articles. ht. J. Epiakmiol. 19, 789-800, 1990. Spielberger Ch. State-Anxiety Inventory: A Comprehen- sive Bibliography. Consulting Psychologists Press, Palo Alto, Calif., 1984. Friedman L., Furberg C. and DeMets L. Fundamentals of Clinical Trials, pp. I-10. PSG Publishing, Littleton, 1985. Bryce R., Stanley F. and Barry Garner J. Randomized clinical trial of antenatal support to prevent preterm birth. Br. J. Obstet. Gynaec. 98, 100-1008, 1991. Stevens A. A randomised controlled trial of community antenatal care in central Birmingham, unpublished. Mosley W. H. and Chen L. An analytical framework for the study of child survival in developing countries. Popular. Develop. Rev. 10, Suppl., 2545, 1984.

Sosa R., Kennel1 J. H., Klaus M., Robertson S. and Urrutia J. The effect of a supportive companion on perinatal problems, length of labor and mother-infant interaction. N. Engi. .I. Med. 303, 597-600, 1980. Elboume D., Richardson M., Chalmers I., Waterhouse I. and Holt E. The Newbury Maternity Care Study: A randomized controlled trial to assess a policy of women holding their own obstetric records. Br. J. Obstet. Gynaec. 94, 612619, 1987. Carpenter J., Aldrich K. and Boverman H. The effectiveness of patient interviews. A controlled study of emotional support during pregnancy. Archs gen. Psychiat. 19, llt%ll2, 1968. Blonde1 B., Breart G. and Llado J. Prevention of preterm deliveries by home visiting midwives: Results of a randomised controlled trial, unpublished. Love11 A., Zander L. I., James C. E., Foot S., Swan A. V. and Reynolds A. St Thomas’ maternity case notes study: Why don’t give mothers their own case notes? In Ciceiy Northcote Trust, pp. l-155. United Medical and Dental School of St Thomas’ Hospital, London, 1986. Jacobson D. Types and timing of social support. J. Hlth sot. Behav. 27, 25&264, 1986. Cohen S. and Syme S. L. Issues in the study and application of social support. In Social Support and Health (Edited by Cohen S. and Syme S. L.), pp. 4-5. Academic Press, New York, 1985. Berkman L. and Breslow L. Health and Ways of Living. The Alameda County Study. Oxford University Press, New York, 1983.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

SSM 36,4--I