The Contribution of Church-Based Networks to Social Care in the Coronavirus Pandemic and Beyond: The Case of Pastoral da Pessoa Idosa in Brazil

Abstract

The Pastoral da Pessoa Idosa (PPI) is a national voluntary movement linked to the National Conference of Bishops of Brazil (CNBB), comprising approximately 2500 volunteers, who offer guidance and support for well-being to more than 170,000 older people throughout Brazil. This paper analyses the results of a data gathering exercise designed to gain an overview of the scope, support needs and international transferability of the project in ‘normal’ times and during the coronavirus pandemic. A telephone questionnaire comprising 21 questions was launched nationally among PPI volunteers over a single week in May 2020. The dataset (n = 3888) was subjected to exploratory statistical analysis and key findings collated around three guiding questions. The central findings were, first, that the composition of the body of volunteers is changing, requiring changes to training and support; secondly, that the model may be exportable to other middle-income countries in which the Catholic Church does not have the same dominance over the culture, as long as a national organising structure is in place; and finally that the speedy and flexible response of the PPI to the coronavirus pandemic suggests that this type of NGO will have a role in response to future national crises.

1. Introduction

Brazil is currently experiencing one of the world’s most rapid rates of population ageing, and with this comes a rise in social care needs. Despite some significant policy initiatives, implementation has stalled because of the difficulty of instituting national structures and the sheer cost of responding to the growing needs (Neumann and Albert 2018). Consequently in Brazil, as in a number of middle-income countries in South and Central America, statutory social care is thinly and patchily distributed, with a concentration of resources and ‘age-readiness’ in larger and more prosperous urban areas (Leitão 2019). Deficits in social care are likely to have their most serious impact on the increasing number of older people, because of their poorer health, the effects of rapid urbanisation, social isolation and relative poverty (Miranda et al. 2016). High levels of social inequality lead to uneven outcomes and have a disproportionate impact upon women, who frequently stop working to care for older people in need of support (Lima-Costa 2018). Furthermore, the rapid growth in the costs associated with an ageing population (such as the state pension scheme projected to cost 18% of GDP by 2060, Neumann and Albert 2018) threaten the whole system of social care provision in the future.

As well as these pressures upon everyday social care, the Covid-19 global pandemic has placed additional strain on support systems. During the period of data gathering for this study (11–18 May 2020), Brazil was firmly in the exponential phase of the outbreak. Daily new cases rose from 6444 on 11/5 to 14,288 on 18/5. Total cases over the same period rose from 169,143 to 255,368, and total deaths from 11,625 to 16,853 (Worldometer 2020). At the time of writing (28 July 2020), Brazil has nearly 2.5 million cases, the second highest number in the world after the USA, and while the death rate has stopped increasing, it has yet to fall (Bhatia 2020). The pandemic has made visible situations experienced by the elderly in Brazil, such as abandonment, ageism, social exclusion, social policies being patchily implemented and a failure to meet the demands of this rapidly growing age group.

Given these pressures on the present and future state provision in Brazil, voluntary networks and organisations could make a significant and increasing contribution to the well-being of older people in the community. The purpose of the study underlying this paper was to examine the role of the main national voluntary organisation supporting the social care of older people, the Pastoral da Pessoa Idosa (PPI). This movement coordinates the work of approximately 25,000 volunteers (Agenti Pastorali, hereafter APs) who, between them, support approximately 170,000 elderly people, providing guidance on self-care, aiming at health promotion, disease prevention, information on public services, personalized support and interchange between APs and isolated older people visited in their homes. This figure represents approximately 1% of the total population of people over 65 in Brazil, based on a population of 210 million, of whom 9.3% are over 65 (Knoema Corporation 2020; Plecher 2020).

Three questions guided the design and execution of this study, and provide the basis for this paper:

• Over a period of rapid growth and geographical expansion, what changes can be identified in the characteristics, location and social circumstances of the volunteers, and how should these be recognised in the planning of their training and support?
• To what extent may this successful model of voluntary social care be transferrable to other middle-income countries with similar economic, geographical and political features; and to what extent is its success predicated on the unique religious ecology of the Catholic Church of Brazil?
• How has the PPI movement responded to Covid-19? Specifically, how is the pandemic affecting the lives of older people, how are the APs mitigating its effects and is the movement capable of providing resilience in the face of this and future disease events?

2. Background

In industrialized, secular economies, religiously-organised health and social care providers often have an uneasy relationship with public policy makers. Concerns may be expressed by policy makers and commissioners of services that, by collaborating with a religiously-motivated organisation, they are laying themselves open to accusations of favouritism or hidden endorsement of a religious agenda; that public resources may be co-opted for a different set of purposes or that a religious organisation may lack transparency and accountability in the way that it distributes resources (Barclay and Hilhorst 2019). Conversely, religious organisations report that they sometimes encounter a conflict between the expectations of their parent body and the expectations of public bodies (Tomkins et al. 2015); that some legislation (for example on equalities, diversity and discrimination) conflicts with historic practices, or that public servants lack insight into the particular needs of religiously-motivated ethnic minorities that have historically lacked access to health and social care (Smith 2004; Lowndes and Chapman 2007). Finally, the emergent ‘minority’ status of religion generally renders it less known and more ‘other’ as a cultural expression (Bruce 2014) and public health officials are not educated in the role of religious communities (Long et al. 2019). Perhaps for these reasons, in many industrialized nations, the significant contribution to the social care of older people offered by religious organisations tends to be overlooked by and poorly integrated into national provision (Kevern 2018).

In many low- to middle-income countries, this division between state and religious provision may be less clearly enforced for a number of reasons. Pragmatically, national provision may be dependent upon the resources of religious communities (and perhaps particularly their access to international aid) to deliver essential health, social care and educational services (Karam et al. 2015): for example, it has been estimated that anywhere between 30% and 70% of health care in Subsaharan Africa is delivered by religious groups (Olivier et al.). Some commentators therefore consider that collaboration between policy makers and religious providers is a key element in delivering global health and multisectoral development (Duff and Buckingham 2015). As Marshall and Smith point out in their study of faith-based organisations in the fight against Ebola, “Strategies to strengthen basic health systems and public health approaches will benefit if they take full account of the on-the-ground presence of religious institutions and draw in an integrated way on relevant disciplines (for example, anthropology, religious studies, and social and behavioural sciences.” (Marshall and Smith 2015, p. e24f). Furthermore, in many countries, the major religious groups are organised in such a way as to enable large-scale coordination of efforts to improve health and social care—a level of organisation that may rival that of national government agencies. In the evocative terminology of Goldenberg (2015) religious groups are ‘vestigial states’, functioning as alternative, past or future social orders. Finally, religious groups have a well-documented capacity to mobilise volunteers (Graham and Haidt 2010) because of the way they promote prosocial behaviour. This means that they typically have human resources which national government bodies cannot match.

The religious, social, economic and political profile of Brazil lends itself to the deployment of religious resources to support health and social care goals. It is a huge and populous country with a wide diversity of geographical and social environments from very large cities to extremely remote rural settlements and a relatively undeveloped health and social care infrastructure. By contrast, the Catholic Church of Brazil is highly organised with 275 individual units of organisation under the oversight of the Conference of Catholic Bishops of Brazil (Cheney 2020). Although formal allegiance to the Church has dropped over the last generation, in the census of 2010, 65% self-declared as Roman Catholic, with ‘no religion’ making up most of the remainder in urban areas and Pentecostals in more remote rural areas (Instituto Brasileiro de Geografia e Estatística-IBGE 2010). The conditions are therefore favourable for a model of social support that has national reach and impact, and a range of such programmes are run under the auspices of the Church (Conferência Nacional dos Bispos do Brasil 2017). The Pastoral da Pessoa Idosa (PPI)) programme is one of these.

The movement has its origins in 1994, in the work of the Brazilian doctor Zilda Arns who, recognising the challenges posed by a rapidly ageing population, initiated a programme to accompany older people through home visits. The programme was launched as a collaboration between all the dioceses of Paraná and the state of Paraná in 1996. After repeated iterations and revisions, training and organisation was standardised on a national basis in 2002 and a manual was developed to support community leaders to work for the well-being of older people. The department, Pastoral da Pessoa Idosa, was approved by the national church in 2004 (Conferência Nacional dos Bispos do Brasil 2005a). After a period of very rapid growth, numbers of volunteers have remained roughly constant since 2010, and the number of older people accompanied has levelled off since 2012 (Barroso 2017).

PPI is a NGO recognised by the Brazilian government but organised through the 212 dioceses of the Catholic Church of Brazil. The Church’s most important contribution is perhaps an organisational model: PPI has a hierarchy modelled on that of the Catholic Church, with a chain of National, Regional, State and Local organisers on local volunteers. The Brazilian Ministry of Health provided financial support for the first six years of PPI’s official existence, but this was discontinued in 2010 and the main source of income has been voluntary donations for the last decade (Tortelli 2013). Perhaps unsurprisingly, this withdrawal of funding coincided with the point at which the growth of PPI slowed markedly. Nevertheless, the main assets of the movement are its volunteers, along with local parishes, communities and individuals who contribute physical space, material resources and sufficient finance to maintain the movement.

APs live in the same community and receive training that lasts an average of 28 hours in total to monitor the pessoas idosas (PIs) in their territory, community or in their building and to ‘accompany’ (acompanhamento) them in response to their needs. They are trained to identify key indicators of frailty, and are encouraged to visit an average of 10 older people monthly (Conferência Nacional dos Bispos do Brasil 2005b), although the figures quoted above suggest the mean figure is approximately 6.8 per AP. Activities of volunteers include practical support (such as collecting shopping or medicines), emotional and spiritual support, but also referrals to local services and the regular monitoring and documenting of the needs of older people. “In different ways, the community leaders of the Pastoral da Pessoa Elderly collaborate for building networks with a focus on aging. In home visits, close ties between cohabitants are established, always respecting dynamics, particularities and family history. It is when they talk about aspects related to health promotion for the accompanied elderly.” (Barroso 2017, p. 441, translated1).

3. Method

A 21-item telephone questionnaire was developed collaboratively by the five authors and piloted with a small ‘expert group’ of researchers and a team of professionals from PPI’s National Coordination. State Coordinators were instructed to disseminate the study and also to encourage the participation of Pastoral Agents throughout the national territory. The interviews were conducted over a 7 day period (11–18 May 2020) by cascading the telephone questionnaire through the national and state coordinating structure to the local coordinators. Local coordinators conducted interviews by cell phone and recorded the responses as they were given. Questionnaire data was collated in an Excel spreadsheet by the three Brazil-based authors as the responses were received and the dataset was finalized on 18th May 2020. Because of the disruption to normal life arising from the Covid-19 measures and the need to gather data in a timely fashion, no fixed sample size was set, although the authors agreed that a sample of 500 would give sufficient breadth for a basic set of descriptive statistics. In the event, 3888 questionnaires were returned, enabling some more nuanced statistical inferences to be drawn from the dataset.

4. Results

Data was returned for 3888 participants, representing approximately 15.5% of the total APs. Data was analysed in Excel and IBM SPSS (version 26).

This was an exploratory questionnaire which did not employ validated scales, delivered ‘in the field’ by telephone using a number of interviewers, and timescale and the limitations of research under restrictions imposed by the pandemic meant it was not possible to check the consistency and reliability of interviewers’ work. For these reasons, the dataset needs to be understood primarily as descriptive statistics which give insight into the broad demographics and activities of the APs. These are summarised in

Table 1. Descriptive statistics2.

	Item	Sample Size (N)	Data Type Continuous (Cs) Ordinal (O) Categorical (Cl)	Key Data
1	Age	3865	Cs	Mean = 55.38 Median = 56 SD = 11.89
2	Sex	3888	Cl	Female: 3567 (91.7%) Male: 321 (9.3%)
3	How many do you live with?	3888	O	Alone 587 (15%) Spouse 1046 (27%) 2 or more 2255 (58%)
4	Health	3888	O	Good 3030 (78%) Regular 844 (22%) Bad 14 (<1%)
5	How many illnesses do you have	3888	Cs	Mean 0.78 Median 1 SD 0.95 Maximum 10
6	How many years have you been an AP?	3888	Cs	<1 year 631 (16%) 1–3 years 1574 (40%) 4–7 years 970 (25%) 8–10 years 417 (11%) >11 years 296(8%)
7	Settlement size	2284	O	>1 m 201 (9%) 500,001 to 1 m 110 (5%) 300,001 to 500,000 174 (8%) 100,001 to 300,000 370 (16%) 50,001 to 100,000 319 (14%) 10,001 to 50,000 318 (14%) 5001 to 20,000 416 (18%) <5000 176 (8%) Rural 200 (9%)
8	How many elderly people are you accompanying?	37823	Cs	Mean: 7.41 Median: 7 SD: 4.05
9	How old are the people accompanied?	3888	O	60–65 = 972 (162 p.a) 66–74 = 1988 (221 p.a) 75–89 = 2800 (200 p.a) 90–99 = 834 (83.4 p.a.) Over 100 years old = 112 (See note below)4
10	What motivates you?	3888	Cl	Follow Christian teachings 2186 (56%) I like spending time and learning with the elderly 2180 (56%) Helping to improve the life quality of older people 2514 (65%) Defending the rights of the elderly 1607 (41%) Being a volunteer enriches me as a human being 2727 (70%) Another Reason 224 (6%)
11	In the last week, how many people were you in contact with?	3888	Cs	Mean = 4.83 Median = 4 SD = 6.10
12	By what means?	3888	Cl	1. Telephone = 2125 (1141, 29%)5 2. Letter = 11 (3, <1%) 3. Email = 16 (0) 4. WhatsApp = 1393 (457, 12%) 5. Other = 910 (576, 15%)
13	In the last week, how did you help people?	3888	Cl	Well-being support (offering company, emotional or spiritual support) = 1047 (327, 8%) General advice = 1032 (106, 3%) Advice about living with coronavirus outbreak = 1948 (797, 20%) Other = 471 (205, 5%) Advice about how to live better during the outbreak = 1165 (160, 3%)
14	Over the last week, how did you offer the elderly people practical support?	3888	Cl	Making a telephone call, letter or email = 0 Shopping for necessities = 782 (311, 8%) Contacting relatives or friends who can supply help = 982 (346, 9%) Referral to another agency, such as the hospital or social services = 780 (232, 6%) Any other form of practical support = 406 (240, 6%) Participated or organized collection of food and hygiene materials or making masks for donation = 1072 (459, 12%) No answer = 1393
15	What effect has the pandemic had on increasing isolation and loneliness?	3888	O	No effect = 31 Small effect = 105 Medium effect = 732 Big effect = 3020
16	What effect has the pandemic had on making people more anxious or confused?	3888	O	No effect = 19 Small effect = 99 Medium effect = 860 Big effect = 2910
17	What effect has the pandemic had on making it more difficult to get practical necessities of life?	3888	O	No effect = 33 Small effect = 124 Medium effect = 881 Big effect = 2850
18	What effect has the pandemic had on making it more difficult to get help from government?	3888	O	No effect = 103 Small effect = 219 Medium effect = 1251 Big effect = 2315
19	Which of these changes have come about as a result of social isolation?	3888	Cl	Stimulated intergenerational interaction = 1063 Solitude = 1644 Stimulated solidarity = 2905 The search for a spirituality = 2713 Encouraged the search for the meaning of life = 2456 Increased domestic violence = 1073 Other = 219
20	What difficulties did you encounter carrying out the PPI monitoring during the Pandemic?	3888	Cl	I didn’t have any difficulties = 1863 Not having electronic equipment (cell phone, computer, etc.) = 625 Not knowing how to use digital technology = 593 Having no strong digital signal in the region where you live = 188 Other = 619

. However, because of the large sample size and the fact that the data was gathered in a narrow time window that reduces the confounding effect of changing circumstances during data gathering, it was possible to conduct some more robust statistical testing to identify where a strong correlation exists between different measures within the dataset and draw statistical inferences. Items that demonstrated a robust statistical correlation are summarised in

Table 2. Item correlations where statistical significance was reached (shaded rows show weak relationship).

Variable 1 (Dependent for ANOVA)	Variable 2 (Factor for ANOVA)	Test Statistic ANOVA (F) Kruskal–Wallis (H)	Significance	Finding
AP age	Number of illnesses suffered	F 53.122	0.000	Older APs have more illnesses
AP age	Number of people lived with	F 208.825	0.000	Younger APs live in larger households
AP age	Number of years as AP	F 84.299	0.000	Younger APs are more recent recruits
AP age	How would you describe the region where you work?	F 11.886	0.000	Smallest settlements have younger APs
How many elderly people are you accompanying?6	How would you describe the region where you work?	F 9.422	0.000	APs in smaller settlements have more PIs
How many elderly people are you accompanying?	Sex	F 16.186	0.000	Males accompany fewer PIs
Total illnesses per APS	Sex	F 16.186	0.000	Males have fewer illnesses
Total illnesses per APS	How many people do you live with?	F 10.714	0.000	APs in larger households have fewer illnesses
How many people do you live with?	How would you describe the region where you work?	H 12.271	0.002	More rural APs live in larger households
How would you describe the region where you work?	Total illnesses per APS	H 11.355	0.003	More recent recruits have fewer illnesses
How many years have you been a Pastoral Agent?	Sex	F18.040	0.000	More recent recruits are more likely to be male
How many years have you been a Pastoral Agent?	How many people do you live with?	F9.815	0.000	More recent recruits live in larger households
How many years have you been a Pastoral Agent?	Total illnesses per APS	F8.737	0.000	More recent recruits have fewer illnesses
How many years have you been a Pastoral Agent?	How many elderly people are you accompanying?	F7.460	0.000	More recent recruits have fewer PIs
How many years have you been a Pastoral Agent?	How would you describe the region where you work?	F11.014	0.000	More recent recruits are likely to live in smaller settlements
In the last week, how many people have you accompanied?	How would you describe the region where you work?	F2.182	0.042	Higher numbers in large (>1 m) and small <50 K settlements

.

For this correlational stage of the analysis, since the dataset comprised a mixture of items scored as continuous, ordinal and categorical data, the strategy was as follows:

• To minimise the risk of false-positive results arising from the random spread of multiple comparisons, the Bonferroni Correction was applied. Eight items from the questionnaire were subjected to correlational analysis (excluding items 4, 9, 10, 12–21) giving a theoretical maximum of comparisons as 36. The Bonferroni correction, gives a revised critical probability level of 0.05/36 = 0.0014 or 0.001 with rounding (Field 2013). Because of the large sample size, where correlations are identified they are typically in the range p < 0.001 and sufficient account of theBonferroni correction can be taken, but where p > 0.001, this is indicated and the finding was treated as of unproven significance.
• Where one of the items being compared comprised continuous data, one-way ANOVA was used. Although it was not possible to establish in each case that the data fulfilled the assumption of normality, ANOVA is robust to such violations (Field 2013, pp. 184, 444). Detailed post-hoc modelling was not attempted because of the limitations of the data, but when significant correlations were identified, inferences were drawn by comparing means for the different categories of the test variable.
• Where one or both of the items being compared was scored as ordinal data, Kruskal–Wallis H was used as the appropriate non-parametric test of relationship. Where a significant correlation was identified, means for the grouping variable were compared visually to determine the nature of the relationship. This procedure was purely for comparative purposes, as means produced by ordinal data lack construct validity, but identified variables that appeared to be varying in comparable ways.
• Item relationships that achieved statistical significance were collated and tabulated as in Table 2 below. Examination of the ways in which correlations clustered then made it possible to develop initial hypotheses regarding the changing characteristics of APs and of their role both over time and in particular in response to the challenges of the Covid-19 pandemic.

5. Analysis

5.1. Observations from Descriptive Statistics

APs are typically in their middle years, and the overwhelming majority are women. Approximately 57% live in households of three or more people, implying continuing responsibilities for parents or children. The majority report themselves to be in good or regular health although half of them by median report at least one long-term health condition. The majority of APs had been part of the programme for 3 years or less, with a declining proportion reporting participation for 4–7, 8–10 and 11+ years. Since the project has grown only slowly over the last decade, this suggests that the difference is due to attrition and that many APs volunteer for only a few years before other commitments or a change of circumstances lead them to withdraw. The way in which this change in numbers is accompanied by a change in composition of the AP community will be examined in more detail in the following section.

The activity of APs is distributed fairly evenly across a range of settlement sizes. An AP, on average, maintains contact with 7.41 PIs, from which it may be inferred that the 25,000 APs collectively monitor and support approximately 185,000 older people—this figure is slightly higher than the average calculated from the national numbers given above. When the age of PIs accompanied is adjusted to take account of the different sizes of the age brackets in the questionnaire, the numbers are comparable for the 66–74 and 75–89 brackets, with lower numbers on either side of this broad band. Little can be inferred from these figures, apart from the self-evident point that younger people are less likely to need help and older people are less abundant in the Brazilian age pyramid (Worldometer 2020).

Motivation of APs is spread evenly across Christian and broadly humanistic concerns, with no major differences, arguing that Christian motivations are not particularly strong additional ones in this case. Thus, in answer to Question 10 (What motivates you?), only 56% included the desire to “Follow Christian teachings” as one of their answers. This is comparable to the number who answered that “I like spending time and learning with the elderly” (56%) and rather fewer than the number answering that they were “Helping to improve the life quality of older people”. Overall, by far the most important driver appeared to be that of personal human enrichment (70%). The only other outstanding feature is that the ‘political’ motivation of defending rights of the elderly is not a major one: human concerns prevail.

Under lockdown conditions in which APs were requested to discontinue personal visiting, there appears to have been a reduction in the average number of contacts (mean 4.83 in the week of interviews) with the primary medium being telephone, followed by WhatsApp. APs who only used one medium overwhelmingly chose telephone, although both WhatsApp and ‘Other’ were well represented. Since 782 people (20%) reported that they had shopped for necessities for an older person in the same week, it is reasonable to assume that such personal interactions represented the majority of contacts in this latter category, notwithstanding PPI advice against personal visits. The same responsiveness is reflected in the proportion of practical help devoted specifically to providing food and protective equipment to deal with lockdown, an activity that engaged 12% of respondents as their sole contribution during that week. Unsurprisingly, when asked what difficulties they had during the lockdown (Q20), a proportion of respondents reported being hampered by the lack of infrastructure or knowledge for operating electronic means of communication. Perhaps more surprisingly, approximately half of all respondents reported no difficulties, and seemed to have adjusted well to the required change in approach.

More than half of the responses concerning interpersonal support offered over the week were directly related to adjusting to the pandemic, which presumably illustrates both the level of concern among PIs and the flexibility with which the APs responded to it. These conclusions are supported by the responses to questions 15–17, which show how seriously APs took the effects of the pandemic on the PIs in their charge. Perhaps surprisingly, respondents were less emphatic in response to the suggestion that the pandemic made it more difficult to access government services, although at this point in the pandemic cycle its effects on this level of provision may have been relatively slight.

The final two questions relate to potential changes in activity during the pandemic. Respondents identified a number of potential positive as well as negative effects which resist further analysis in questionnaire form.

5.2. Observations from Comparative Statistics

Seven of the eight items tested demonstrate some statistically-significant relationship to at least one other item. The most widely correlated result was for the variable “How many years have you been a Pastoral Agent?” which was significantly correlated with six other items. Further examination of the dataset demonstrates that there has been a notable change in the characteristics of APs over the course of the project’s implementation. More recent recruits are disproportionately younger, healthier, living in larger households, more present in smaller settlements and rural areas and include more males. They are also actively accompanying fewer PIs, which may be reflecting the time it takes to build durable supportive relationships.

Settlement size seems to be a key variable here, displaying strong correlations with three other variables as well as weaker ones with three more. As noted, small settlements tend to have younger APs but, somewhat contrary to the relationship noted above between time as an AP and number of PIs, these APs have wider networks than average. APs living in smaller settlements tend to live with more people and have fewer illnesses, although given that age is strongly correlated with these variables the fact that they are younger may explain these characteristics.

The APs with most PIs tend to be women who have been APs for a number of years and/or come from smaller settlements. However, these characteristics did not show a strong statistical relationship with the APs’ level of activity during the last week, when the only (weak) statistical conclusion was that APs based in large (>1 m) and small <50 K) settlements were more active than those in intermediate-sized towns and cities. This may reflect the increased ability of APs in large centres to communicate with and coordinate help to PIs as well as the flexibility and commitment of APs in rural areas, but further work needs to be done to test this theory.

The relationship between the test variables can be visualised as in Figure 1 (broken lines signify weak relationships):

6. Discussion

This paper began with three questions which we intended to address through this broad-ranging questionnaire on the characteristics and activity of APs. The analysis above of descriptive statistics does not enable rigorous inferences to be drawn, but suggests the following broad conclusions that need to be followed up by further study.

6.1. Characteristics of APs

The picture of a ‘typical’ AP emerging from the data is of a woman in her middle years, in fairly good health, living in a household of three or more people and therefore likely to have continuing family responsibilities. As the programme has undergone rapid growth, the preponderance of APs are relatively recent recruits who are more likely to be younger, healthier, male (although males are still a small minority) and living in larger households. As the programme has expanded out geographically from its original urban centres, an increasing proportion of them are from smaller and more remote communities, although the majority are still from settlements of at least 50,000 people.

These findings suggest some directions for future policy developments and administration of PPI within Brazil:

• Since the characteristics of recruits are changing, training and support may need to take account of this. For example, recruits are younger, healthier and potentially more active, but may therefore have more commitments and less time to devote to training. The increasing proportion from rural areas and small settlements may have more difficulty accessing training events or support from regional coordinators. Finally, in more rural areas APs are likely to be working in relative isolation, and the development of the training and support programme needs to take this into account.
• The average number of PIs being supported is encouraging at 7.41, suggesting that the programme has a substantial contribution to make to the wellbeing of older people across Brazil. However, this figure conceals a wide variation in the number of older people supported. Only three respondents said that they were not currently supporting any PIs, but further work to uncover the sources of the variation, and whether improved support would benefit those who are not currently very active, is advisable.

6.2. The Reach and Exportability of PPI

It is noteworthy that the motivation is not specifically or exclusively Christian for a large number of respondents, but is equally likely to be described along generically humanist lines. Personal enrichment and a desire to help older people were as significant as religious commitment in motivating APs. This suggests that although PPI undoubtedly benefits from the distinctive teaching and organisational infrastructure of the Catholic Church in Brazil, it might be susceptible to adaptation to other contexts in which the religious ecology is not the same, as long as appropriate organisational and support structures can be maintained through other means. These may be provided by alternative religious structures such as Muslim charities in Indonesia (Kailani and Slama 2020) or secular political or social movements such as HelpAge India (Liebig and Rhajan 2003; Jha and Mohapatra 2017).

The question of organisational structure, motivation and support is clearly critical in this context. The Catholic Church of Brazil is exceptional in being a non-governmental organisation that has national (if variable) reach across the whole range of settlements from large metropolitan areas to small villages; and although formal allegiance has declined in recent decades it retains widespread support as a social institution (Estela de Sousa Pinto 2016; Sousa 2018). In addition, it has a rare degree of internal organisation and communication which enables large-scale motivation, training and coordination of volunteers.

The effect of this organisational sophistication is apparent in the data gathering process for the current study. The accumulation of 3888 responses via telephone interview in the space of a single week is a consequence of a clear and committed chain of state, regional and local coordinators. Clearly, this degree of integration is a major driver for the reach, impact and sustainability of PPI as a movement and may well be a more significant determinant of its success than the involvement of the Catholic Church per se. A reasonable inference is that PPI may well be ‘exportable’ to other countries with fragile governmental infrastructures for health and social support in the community, but only if an alternative structure for coordinating and supporting activity can be located.

6.3. Pastoral Support in a Time of Covid-19

The analysis of AP characteristics suggests that there are two distinct subgroups who are at risk of contracting and spreading Covid-19 in different ways. The first is comprised of older APs who live alone or with a spouse and suffer from multiple health conditions: for these individuals, an appropriate response to the pandemic would be to cease all activities that may expose them to risk of serious illness or death. The second is comprised of younger, relatively healthy individuals who live in larger households: in this case, the risk is that they may contract and spread the virus through their multiple close contacts with others, and so endanger the people they are accompanying. With hindsight, the PPI national office made a timely and appropriate decision when they asked APs to cease all personal visits in March, although the proportion of APs who listed ‘shopping for necessities’ among the forms of practical support they had recently provided suggests that the injunction was interpreted flexibly.

Data collection took place in a week during which anxiety about Covid-19 was high as the numbers rapidly multiplied; in addition, the PPI central organisation has asked APs to stop personal visiting on 23 March, forcing them to find other means of supporting PIs. This is despite the PPI’s main mission as being to conduct home visits to the elderly (Conferência Nacional dos Bispos do Brasil 2020) The dataset therefore provides a snapshot of the way APs adjusted to an unexpected restriction on their activities.

Although there was some reduction in the level of contact during the week under review (mean 4.83 in the week of interviews compared with 7.41 under normal conditions), this still represents a significant level of support. The limiting factor here may have been difficulty using electronic means of communication: although APs adapted by making contact by telephone or WhatsApp, just over half reported some difficulty in using these media.

Most striking was the way in which APs redirected their energies towards dealing with the challenges of the lockdown, form the offering of practical help (food, protective equipment) to Covid-related advice and guidance, to organising or participating in volunteer groups to make PPE or collect and distribute food and protective materials.

It is not possible to quantify the contribution of PPI to population resilience to the stresses of the pandemic. However, there is abundant evidence that it contributed to mobilising a community response, as well as continuing its supportive work among older people by adapting its practices in a timely and flexible way. This raises the question of whether, when public health departments start to consider how to prevent a resurgence of this pandemic or how to deal with a subsequent one, the resilience-building activity of organisations such as PPI should be taken into account and actively developed. As the pandemic has demonstrated, middle-income countries with poorly-resourced statutory social care are liable to be quickly overwhelmed by large-scale shocks such as pandemics or extreme climate events, and such organisations may have the flexibility and energy to respond rapidly at the level of individual communities.

7. Conclusions

This paper raised three questions regarding the scope and activity of organisations such as PPI in Brazil and internationally.

• What changes can be identified in the characteristics, location and social circumstances of the volunteers?
• To what extent may this successful model of voluntary social care be transferrable to other middle-income countries with similar economic, geographical and political features?
• What has the PPI movement contributed to the response to Covid-19?

The forgoing discussion has gone some way to answering them, although further and more focussed study will be required to add detail to this initial survey. On the basis of the analysis above, it is possible to conclude three things, each of which have implications for policy making in their respective arenas.

First, on the level of national policy, that PPI is a dynamic movement which, as it matures, needs to keep revising its training and support to take account of the changing nature of the volunteers and the demands upon them. It is making a substantial contribution to the social care of vulnerable and isolated older people that will only become more important in the face of a fast-growing demographic challenge. However, in order to maintain and support a sustainable network of volunteers, it may be necessary to adjust resource provision, training and the development of intergenerational dialogue and support to take account of the changing volunteer profile.

Secondly, at the transnational level, there is some basis for confidence that this model may be transferrable to other middle-income countries with similar social geography and demographic issues. Despite the prevalence of the Catholic Church, it does not appear that Brazil’s unique religious ecology is the determining factor in the success of PPI. However, the success of this model does appear to be dependent on a well-developed organisational structure to provided rapid, efficient and authoritative communications and support. In the absence of comparable religious structures, secular analogues would need to be identified if this model is to successfully take root in other national contexts.

Finally, at the global level, this type of movement has shown itself to be agile, committed and effective in mounting a response to an unexpected national crisis: features which render it indispensable in a national context where, for whatever reason, the government response has been slow and inconsistent at a number of levels. There is some concern at the global level for how low- and middle-income countries with relatively undeveloped infrastructure and social support will fare if, as expected, the 21st century is likely to deliver a series of population-level shocks: not just pandemics as a result of increased global mobility, but extreme weather events arising as a result of global warming and critical shortages of basic commodities such as water and raw materials (Kikwete et al. 2016). Notwithstanding the reluctance of policy makers to engage with social movements that use religious organisations as a vehicle, these remain the most reliable structures in many low- and middle-income countries for enabling this strategy of global resilience.