Article Text

Impact of conditional and unconditional cash transfers on health outcomes and use of health services in humanitarian settings: a mixed-methods systematic review
1. Kim Robin van Daalen1,
3. Rosemary James3,
4. Henry Charles Ashworth4,
5. Parnian Khorsand5,
6. Jiewon Lim6,
7. Ciaran Mooney7,
8. Yasmeen Khankan8,
10. Isla Kuhn10,
11. Helene Juillard11,
12. Karl Blanchet12
1. 1Cardiovascular Epidemiology Unit, Department of Public Health & Primary Care, Cambridge University, Cambridge, UK
2. 2UCD Centre for Interdisciplinary Research, Education and Innovation in Health Systems, School of Nursing, Midwifery and Health Systems, University College Dublin, Dublin, Ireland
3. 3University Hospitals of North Midlands NHS Trust, Stoke-on-Trent, Stoke-on-Trent, UK
4. 4Harvard Medical School, Boston, Massachusetts, USA
5. 5Women in Global Health, Washington, District of Columbia, USA
6. 6School of Medicine, NUI Galway, Galway, Ireland
7. 7Northern Ireland Medical and Dental Training Agency, Belfast, Antrim, UK
8. 8Department of Biology, Siena Heights University, Adrian, Michigan, USA
9. 9Kabul University of Medical Sciences, Kabul, Afghanistan
10. 10Medical Library, School of Clinical Medicine, University of Cambridge, Cambridge, Cambridgeshire, UK
11. 11Geneva Centre of Humanitarian Studies, University of Geneva, Geneva, Geneva, Switzerland
12. 12Global Health Development, University of Geneva Faculty of Medicine, Geneve, Switzerland
1. Correspondence to Ms Kim Robin van Daalen; krv22{at}cam.ac.uk

## Abstract

Background Cash transfers, payments provided by formal or informal institutions to recipients, are increasingly used in emergencies. While increasing autonomy and being supportive of local economies, cash transfers are a cost-effective method in some settings to cover basic needs and extend benefits of limited humanitarian aid budgets. Yet, the extent to which cash transfers impact health in humanitarian settings remains largely unexplored. This systematic review evaluates the evidence on the effect of cash transfers on health outcomes and health service utilisation in humanitarian contexts.

Methods Studies eligible for inclusion were peer reviewed (quantitative,qualitative and mixed-methods). Nine databases (PubMed, EMBAS, Medline, CINAHL, Global Health, Scopus, Web of Science Core Collection, SciELO and LiLACS) were searched without language and without a lower bound time restriction through 24 February 2021. The search was updated to include articles published through 8 December 2021. Data were extracted using a piloted extraction tool and quality was assessed using The Joanna Briggs Critical Appraisal Tool. Due to heterogeneity in study designs and outcomes, results were synthesised narratively and no meta-analysis was performed.

Results 30 673 records were identified. After removing duplicates, 17 715 were double screened by abstract and title, and 201 in full text. Twenty-three articles from 16 countries were included reporting on nutrition outcomes, psychosocial and mental health, general/subjective health and well-being, acute illness (eg, diarrhoea, respiratory infection), diabetes control (eg, blood glucose self-monitoring, haemoglobin A1C levels) and gender-based violence. Nineteen studies reported some positive impacts on various health outcomes and use of health services, 11 reported no statistically significant impact on outcomes assessed and 4 reported potential negative impacts on health outcomes.

Discussion Although there is evidence to suggest a positive relationship between cash transfers and health outcomes in humanitarian settings, high-quality empirical evidence, that is methodologically robust, investigates a range of humanitarian settings and is conducted over longer time periods is needed. This should consider factors influencing programme implementation and the differential impact of cash transfers designed to improve health versus multipurpose cash transfers.

PROSPERO registration number CRD42021237275.

• child health
• mental health & psychiatry
• nutrition
• public health
• systematic review

## Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information. Not applicable.

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

## Statistics from Altmetric.com

### Key questions

• Previous studies have demonstrated the benefits of cash transfer interventions in low and middle-income countries on mitigating the health impacts from climate change, improving nutrition and advancing maternal health when markets are functional and quality services are available.

• Cash and voucher assistance amount to over US$6 billion in humanitarian aid, with cash transfers accounting for almost three-quarters of this aid. • However, the extent to which cash transfers impact health in humanitarian settings remains largely unexplored. #### What are the new findings? • To our knowledge, this is the first mixed-methods systematic review exploring the impact of conditional and unconditional cash transfers specifically on health outcomes and usage of health services in a humanitarian setting. • Health outcomes assessed in studies largely focused on diet and nutrition, mental and psychosocial health and self-reported general well-being. • Nineteen studies reported some positive impacts on various health outcomes and use of health services, eleven reported no statistically significant impact on outcomes assessed and four reported potential negative impacts on health outcomes. ### Key questions #### What do the new findings imply? • Although our systematic review suggests that there may be a positive impact of cash transfers on health outcomes in humanitarian settings, high-quality empirical evidence, that is methodologically robust, investigates a range of humanitarian settings, and is conducted over longer time periods is needed. • Specific attention must be given to the intended/expressed purpose of grants, the actors involved in designing and implementing cash transfers, and the factors that affect implementation such as local involvement and context-specific considerations. ## Introduction A record 274 million people are expected to need humanitarian assistance in the year of 2022,1 due to increasing extreme weather events, conflict and pandemics. People affected by humanitarian crises have higher rates of poor health outcomes due to increased vulnerability to violence, infectious diseases, food insecurity and chronic diseases. Collectively this picture calls for impactful, innovative solutions able to address a complex range of health challenges including both communicable and non-communicable disease prevention and treatment, water and sanitation, nutrition, access to sexual and reproductive health services and mental health and psychosocial support.2 3 One possible intervention to improve outcomes across these health challenges is the provision of cash transfers. Cash and voucher assistance account for 19% of international humanitarian aid, amounting to over US$6 billion in 2020.4 Cash transfers compose 71% of this aid and are an increasingly common method for providing direct assistance to populations in humanitarian settings.1 4 Cash transfers, which are currently being used in COVID-19 pandemic responses across different settings as well as previously being provided during the Ebola epidemics,5 work by providing individuals with cash to access services or goods. In contrast to vouchers and in-kind assistance, cash transfers are by definition unrestricted in usage and provide recipients with physical currency or e-cash to spend. These can be conditional, where there is a prerequisite activity or obligation that the recipient must fulfil in order to receive assistance, or unconditional, where transfers are provided without the recipient having to do anything to receive the assistance.6

Previous studies and systematic reviews have demonstrated the benefits of cash interventions on mitigating the health impacts from climate change, improving nutrition and advancing maternal health if markets are functional and quality services are available.7–9 Likewise cash transfers can improve outcomes by increasing consumption of nutritious foods as well as access to preventative care and vaccinations.10–12 The effect of cash transfers on health systems is also thought to be beneficial by increasing incentives to seek care or lessening financial barriers to access care.10 Additionally, cash transfers are often preferred by recipients and can offer certain benefits over the direct provision of food, goods or services such as cost-effectiveness, rapid and flexible implementation and transparency.13 14 Cash may also empower recipients by providing autonomy to spend money based on personal need and improve economic growth and stability by enabling the purchase of goods and services freely on the local markets.15 16 While cash transfers offer this flexibility, their effectiveness is also influenced by when and how they are provided (eg, delivery approaches).17 How these interventions are designed and implemented influences the impact of the aid; some of these implementation factors include who is targeted as recipients and their value for money, the state of local markets and infrastructure, available resources, community acceptance and risks relating to security in the setting.14 18 19

To date, multiple systematic reviews have explored the impact of cash transfers on human health or well-being.10 14 19–24 The findings of these reviews have been relatively consistent, suggesting the beneficial impact of cash transfers on the health outcomes of individuals and communities. For example, reviews have identified the utility of conditional cash transfer programmes in improving access to preventive services, while also acknowledging the potential influences of other components.10 20 However, it is important to consider the unique contexts and challenges of populations in humanitarian crises over the general population. Humanitarian settings may be characterised by disrupted health systems and supply chains as well as political and economic instability. People living in humanitarian settings are, therefore, more vulnerable to illness, and subject to worse access to health services and care, and health protection, than the average population. These are contextual factors that need to be considered in the delivery of any health intervention that may not necessarily be addressed in a more static, stable community. Only one systematic review22 specifically examined the effect on health outcomes and the utilisation of health services in humanitarian settings. This review, conducted 6 years ago, focuses solely on unconditional cash transfers (UCTs) and only synthesised three low-quality studies focused on drought contexts. Therefore, in this, we assess the impact of conditional and UCTs on health outcomes and utilisation of health services in humanitarian settings to provide evidence informing future cash transfer interventions and humanitarian response.

## Methodos

This systematic review protocol was prospectively registered on PROSPERO (CRD42021237275). Findings were reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (online supplemental table 7).7

### Definitions

Cash transfers were defined following The Cash Learning Partnership definition as cash payments (physical currency or e-cash) provided by formal and/or (eg, government, non-governmental organisations) informal (eg, hawala—an informal money transfer system largely used in the Middle East and South Asia)25 institutions to recipients that enable them to meet minimum life necessities.6 Unconditional (transfers are provided without the recipient having to do anything to receive the assistance), conditional (there is a prerequisite activity or obligation that the recipient must fulfil in order to receive assistance, eg, attendance of an educational course) and short-duration and ongoing cash transfer interventions were included. Cash transfers are unrestricted, in contrast with (food) vouchers and in-kind assistance.6

The primary outcome(s) were changed in health outcomes, including mortality, morbidity, (mal)nutrition, mental health and well-being and acute or chronic disease status. Secondary outcome(s) were the utilisation of health services, measured by the frequency of visits or percentage of population eligible for a service attending the service.

A humanitarian setting was defined as an event or series of events that present a critical threat to the health, safety, security or well-being of a community or other large group of people, usually over a wide geographic area. Three types of humanitarian crises were specified: man-made crises (eg, civil and inter-state war, armed conflict, genocide), natural disasters including hydrological (eg, floods, avalanches), geophysical (eg, earthquakes, volcanic eruptions, earthquakes), climatological (eg, droughts, wildfires), meteorological (eg, storms, cyclones), biological events (eg, pandemics, epidemics, plagues) and complex emergencies (emergencies resulting from a combination of both natural and man-made causes).26–28

### Search methods and information sources

We searched nine electronic databases (PubMed, EMBASE via Ovid, Medline via Ovid, CINAHL via EbscoHost, Global Health via EbscoHost, Scopus, Web of Science Core Collection, SciELO and LiLACS) without restriction of language and without a lower bound time restriction for articles published through 24 February 2021. An updated search was conducted to include articles published during the COVID-19 crisis through 8 December 2021. Using a combination of free-text terms and subject headings, we used vocabulary related to ‘cash transfers’ and ‘humanitarian settings’. The full-search strategy, developed with a librarian/information specialist, is provided in online supplemental table 1. We conducted forward and backward screening of all articles in the full-text screening phase as well as relevant publications (eg, reviews, opinion pieces) to find any additional studies fitting the inclusion criteria. We also searched Google Scholar to find additional publications.

### Study selection

Eight researchers were involved in the study selection and extraction. After removing duplicates using Endnote, abstracts and titles were screened independently by two researchers according to the selection criteria by using the software Rayyan (https://rayyan.ai/). Studies satisfying the inclusion criteria were retrieved and screened by full text. Conflicts between two authors screening the same studies were resolved among authors until consensus was reached at both stages. A third arbiter was involved when consensus could not be reached. We included primary peer-reviewed quantitative, qualitative and mixed-method studies that either (1) reported on the effect of cash transfers on health outcomes in humanitarian settings or (2) reported on the effect of cash transfers on healthcare utilisation in humanitarian settings. We excluded studies that were (1) non-human studies, (2) conference proceedings or secondary studies (eg, reviews), (3) lacking a full text. In order to capture rigorous data of peer-reviewed studies and evidence, we did not include grey literature (such as non-governmental organisation (NGO) reports). As our research team is fluent in a range of languages (including Arabic, Dutch, English, Farsi, French, German, Pashto, Persian, Spanish, Urdu), we did not exclude any articles based on language; non-English full texts were translated or reviewed by a native or fluent speaker.

### Data extraction and study quality assessment

Data from included studies were independently extracted in duplo using a pretested extraction tool. Any discrepancies between authors extracting the same studies were discussed until consensus was reached. Likewise, a third arbiter was involved when consensus could not be reached. The following information was extracted for each study: author, year, study title, study design, study population, participant demographics (eg, age, country), type of humanitarian setting, sampling and recruitment procedures, total number of participants, outcome (use of health services, health outcomes), outcome(s) ascertainment, type of cash transfer, percentage/number of individuals reporting the outcome, association measures with summary estimate and 95% CI. An open field to record any additional relevant information was available. The quality of individual included studies was assessed using the Joanna Briggs Institute (JBI) critical appraisal tool to explore methodological quality of the synthesised knowledge.8

### Statistical and thematic analysis

Due to the heterogeneity of the included studies (in type of cash transfer, outcome and setting), the quantitative data were descriptively synthesised, and no meta-analysis was performed. Studies with qualitative data underwent additional qualitative thematic analysis. Authors independently used inductive analysis to develop and agree on a codebook. This codebook was then applied to all qualitative studies by two independent authors (inter-rater reliability kappa score 0.96).

### Patient and public involvement

Due to the nature of this study (systematic review), no patients or public were involved in conceptualising or conducting the study.

## Results

### Characteristics of included publications

We identified 30 673 records from the databases. After removing duplicates, 17 715 records were screened by title and abstract and 201 were screened in full text (figure 1). In total, 23 articles were included in this review and their summary characteristics are reported in table 1.29–51Table 2 provides summaries of the main results of the included studies. Included studies were conducted in 16 countries: Niger (n=3),31 32 46 Somalia (n=3),33 34 39 Afghanistan (n=2),35 37 Jordan (n=2),41 48 Lebanon (n=2),49 50 Yemen (n=2),45 47 Palestine,29 Democratic Republic of Congo,30 Syria,36 Cameroon,37 Uganda,38 Bangladesh,40 Ecuador,42 Mexico,43 Kenya44 and Togo.51 Eighteen of the studies were quantitative,30–35 38 39 42–51 two studies were qualitative37 41 and three studies were mixed methods.29 36 40 This included several study designs: randomised control trials (n=7),30 38 42 43 45–47 cohort studies (n=5),31 33 34 36 49 quasi-experimental studies (n=6),29 32 40 44 48 50 non-randomised control trials (n=2),35 39 a cross-sectional study51 and qualitative study designs using in-depth interviews (n=5),29 36 37 40 41 focus group discussions (n=4)29 37 40 41 or observations (n=1).29 The number of participants included in studies ranged from 140 individuals41 to approximately 24 000 households.43 Though no time restriction was applied to the search, all studies were published after 2010, with the vast majority published since 2018.32–37 39–41 44 45 47 Studies were conducted as early as 1998–2000 43 and as recently as July 2020 (during the COVID-19 pandemic).51

Figure 1

Flow diagram of included studies.

Table 1

Summary characteristics of included studies

Table 2

Summary of main results and conclusions

Cash transfers were implemented by governments (n=4)29 35 43 51 and humanitarian agencies/NGOs (n=19)30–34 36–42 44–50 such as the World Food Programme, United Nations Children’s Fund (UNICEF), UN High Commissioner for Refugees (UNHCR), and the International Rescue Committee (IRC). However, studies did not include sufficient detail describing the logistics/involvement regarding which actors served as programme implementers versus delivering cash. While only five studies reported specifically on health service utilisation,35 48–51 a range of health outcomes were investigated across the studies: nutrition-related outcomes (n=13),30–34 36 39 42–47 psychosocial and mental health (n=6),29 31 36 38 40 41 general/subjective self-reported health and well-being (n=3),30 37 acute illnesses (eg, diarrhoea, respiratory infection),50 diabetes control (eg, medication adherence, blood glucose self-monitoring, haemoglobin A1C (HbA1C) levels, Body Mass Index (BMI), blood pressure)48 or gender-based violence.36 Although not part of our inclusion criteria or our main outcome of interest, included studies also reported on expenditure on medicine/paying for healthcare (n=4)29 30 40 49 and self-reported access to health or medication (n=2).41 49 Analysis of qualitative data highlighted additional insight relating to mental health, access to health and the challenges relating to cash transfers (table 3).

Table 3

Qualitative data analysis

### Quality assessment of individual studies

Though no article was excluded from the review synthesis based on quality, the results of the individual study quality appraisals are available in online supplemental table 3-6. Most studies demonstrated reasonable methodological quality, with quasiexperimental studies,29 32 40 44 48 50 indicating the highest quality assessment within the JBI checklist components for quasi-experimental studies. It was unclear whether strategies to address incomplete follow-up were utilised for any of the cohort studies,31 33 34 36 49 and confounders were either not clearly stated or were not included. The latter was similarly observed for the only cross-sectional study included.51 Two of the control trials35 39 were not randomised and, due to the nature of the intervention (receiving cash vs another type of assistance/no assistance), participants could not be blinded in the control trials.30 35 38 39 42 43 45–47 Furthermore, less than half of the qualitative studies located the researchers culturally and theoretically,36 37 and, therefore, did not address the role and influence of the researchers on their findings. It was also unclear for several qualitative studies whether obtained ethical approval.40 41

### Humanitarian setting characteristics

The included settings ranged across different types of humanitarian crises. Broadly, they can be divided into man-made disasters (n=17)29 30 33–39 41 42 44 45 47–50 and natural disasters (n=8).31–34 40 43 46 51 More specifically, these settings included (ongoing) conflict or civil war (n=10),29 30 33–38 45 47 refugee settings and internal displacement camps (n=7),39 41 42 44 48–50 drought (n=3),33 34 43 food crises (n=5),31–34 46 flooding (n=1),40 and the COVID-19 pandemic (n=1).51 Two studies reported on a combination of human conflict and drought in Somalia.33 34 The internal displacement camps were in Somalia39 and Ecuador,42 while refugee settlements based in Kenya hosted refugees escaping civil conflicts in Southern Sudan and Somalia.44 Both Jordan and Lebanon supported Syrian refugees escaping civil conflicts.41 48–50

### Cash transfer characteristics

The majority of the cash transfer programmes examined was unconditional (n=17).29–31 33 34 36 38–42 45 46 48–51 Although five studies included CCTs, one did not explicitly state the conditions to be met.37 43 44 47 48 The four other CCTs had different conditions: (1) for women to deliver their baby at a health facility,35 (2) attendance of sessions on child and infant feeding/care practises for mothers,32 (3) attendance of nutritional training sessions and compliance with child monitoring and treatment for malnutrition47 or (4) quarterly group education sessions on diabetes control, community health worker (CHW) home visits and the provision of receipts to prove use of appropriate services.48 For two studies, it was unclear whether the cash transfers were conditional or unconditional. Cash was usually provided in the short-term over several months (n=12)30–34 36 37 39 42 45 46 50 and ranged in amount; for example, US$96 total over 4 months,46 US$130 over 7 months,30 US$420 over 5 months,39 US$296 over 6 months.31 Fewer interventions utilised one-time cash transfers (n=3)35 37 40 or cash transfers that were provided for more than 12 months (n=2).47 50 The monetary value (n=5)29 41 43 50 51 and duration of cash transfer programmes (n=7)29 41 43 44 48 49 51 were not clearly provided in some studies. While all included cash transfers were used in humanitarian contexts, they varied in purpose: reducing (intergenerational) poverty and economic hardship (n=4)29 38 43 50; increasing households’ access to basic food and non-food needs (n=6)30 37 41 44 48 49; improving food security or preventing (child) acute malnutrition (including reducing child wasting or promote child weight gain) (n=9)30–34 42 45–47; preventing sale or loss of household assets (n=2),31 40 improvement of the use of maternal and child services (n=1)35; improving the ability to take preparatory early action ahead of a natural disaster (n=1)40; reducing the shocks caused by the COVID-19 pandemic51 and prevent negative impacts on health and livelihood (n=1).40 In two studies, the purpose of the cash transfer programme was not clearly defined.36 39

### General health and well-being

Only two studies reported on general health and well-being. Internally displaced persons in Cameroon (US$75–US$100 monthly over 5 months, US$43–US$84 monthly over 6 months) and Afghanistan (one-time US$80–US$198 over 2 months) reported general improvements in health, nutrition and housing after receiving multipurpose cash (MPC) transfers through an International Red Cross programme.37 Similarly, evidence from the Kalobeyei settlement in Kenya indicates that cash transfers (US$14 per person per month) positively impacted nutrition, subjective well-being and independence from aid.44 One study on UCTs (US$130 over 7 months) in the Democratic Republic of Congo presented contrasting evidence and suggested that prevalence of illness and deaths was similar between cash and voucher group.30

### Diet and nutrition-related outcomes

The most commonly investigated health outcomes were related to diets and nutrition.31–34 36 39 42–47 This was assessed in studies using a range of different metrics: the household hunger scale,33 dietary diversity (including, eg, minimum dietary diversity for women,33 34 39 42 44 household dietary diversity score (HDDS) and children’s dietary diversity),31 32 43 45–47 mid-upper arm circumference (MUAC),31–34 39 46 weight for height Z-score (WHZ),31 32 39 46 height for age Z-score,47 food consumption score (FCS),39 42 44 45 caloric intake,42–45 minimum acceptable diet,34 meal frequency,31 33 34 weight gain,32 weight gain velocity32 and the household food insecurity access scale.36 44 Six studies characterised the incidence and prevalence of acute malnutrition as measured by a WHZ < −2, an MUAC <125 mm or the presence of bilateral pitting oedema.31–34 39 46 The impact of cash transfers on nutrition varied. Different outcomes were reported based on the comparison group (eg, in-kind, voucher, no assistance), setting and programme. Studies presented both positive and null effects, rather than a consistently positive or negative effect as further described in the upcoming section.

Assessment of the impact of emergency CCTs (US$250 over 3 months) on the nutritional status of children in Niger found that the intervention was associated with a 1.27 kg overall weight gain (p value <0.001) and 1.82 greater increase in WHZ (p value <0.001) compared with the concurrent control group that did not receive the cash transfer. Furthermore, the odds of having acute malnutrition were 25 times higher for the comparison group.32 Previous evidence in the same setting indicated that among households targeted by emergency UCTs (US$296 over 6 months), diet-related factors and food expenditure for children were not associated with reduced risk of acute malnutrition.31 Likewise, evidence suggested that short-term emergency UCTs (monthly transfer of US$76 for 3 months) yielded significant improvements in food security in the Raqqa Governorate.36 In contrast, another monthly UCT (US$84 for 5 months) combined with a once-only distribution of a non-food-items kit and provision of piped water in Somalia found conflicting results when assessing its impact on acute malnutrition among children 6–59 months. Adjusted for age and sex, the intervention did not appear to reduce risk of acute malnutrition (HR 0.94, 95% CI 0.51 to 174) but did seem to increase the child dietary diversity score by 0.53 (95% CI 0.01 to 1.05).39 In Yemen, increases in dietary diversity (for both children and adults) were observed in the intervention group receiving CCT compared with the control group.47

Despite this, several studies reported positive impacts of cash transfers on mental health (n=4), despite this not being the intent of the cash transfer. Several women in the Raqqa Governorate reported in qualitative interviews that their levels of stress, as well as feelings of humiliation and shame, were reduced in the period of cash assistance delivery. ‘The cash we received maintained our dignity and met our needs. I don’t live like a queen because of the aid, but it is good’, mentioned a divorced woman living with her in-laws.36 Similarly, a study in Bangladesh indicated that forecast-based UCTs (US$60 one-time) reduced psychosocial stress during and after the flood when compared with a group that did not receive cash.40 An UNHCR UCT programme in Jordan for Syrian refugees reduced stress and anxiety among beneficiaries. One third of the respondents in the study indicated their mental well-being had improved and that stress related to inability to pay their rent was alleviated. One of the male beneficiaries reported that, ‘receiving the UNHCR cash transfer changed our life on all moral and financial aspects, I no longer worry about rent and it eased pressure on the entire family’.41 This was consistent with earlier findings from the Palestinian National Cash Transfer Programme on Gazan children.29 ### Acute childhood illnesses A singular study explored the impact of cash on acute illnesses, comparing discontinued recipient households, short-term recipient households, long-term recipient households and non-beneficiary households of MPC transfers. In children under five, short-term and long-term participants suffered less acute illnesses than non-recipients. This finding was further confirmed with lower incidence of specific acute diseases such as diarrhoea and respiratory infections in recipient children versus non-recipient children.50 ### Health service utilisation, access to health and health expenditure The five studies on health service utilisation in this review focused on overall healthcare utilisation,51 maternal and newborn care,35 needed and accessed primary healthcare (PHC),50 diabetes service utilisation,48 care-seeking behaviour for children and chronic or acute illness among adults.49 While almost all positive, there were mixed results relating to what degree CCTs improved healthcare utilisation. Two studies reported on how CCTs affected maternal and child healthcare utilisation. Mothers who received US$15 if they delivered a child at a health facility in Afghanistan reported an increase in both maternal and newborn service usage and CHW home visits. However, only the increase in antenatal care (ANC) visits was statistically significant (adjusted mean difference 45% (95% CI 0.18 to 0.72), p value 0.004). In contrast, the mean difference in postnatal care (PNC) visits, CHW home visits and facility deliverywere not statistically significant.35 Furthermore, MPC recipients (~173.5 US\$ per family per month in Lebanon) reported a 19.3% (CI: 7.3,31.20%; p = 0.002) greater increase in care-seeking behaviour for childhood illnesses compared with controls, and a significantly smaller increase in child hospitalisations among MPC recipients than controls (DiD −6.1%; p = 0.037).49 Likewise, a lower likelihood for needing PHC and a higher likelihood of seeking PHC when needed was observed among children under 5 years old from MPC recipient families.50

A study conducted in a population of Syrian refugees compared the effects of CCTs alone, health education alone and CCTs plus health education on healthcare utilisation for diabetes.48 The study found that the CCT plus health education group had the most significant increases in monthly medication spending (13.6%, p < 0.001) and outpatient diabetes visits (25.3%, p< 0.001). Additionally, the health education only group experienced a decreased overall spending on diabetes care (− 18.7%, p= 0.001).48 This study overall concluded that CCTs were most effective when combined with the health education intervention.

A study on survey data in Togo showed that a government CCT programme, the NOVISSI scheme, improved healthcare utilisation during the economic hardship from COVID-19.51 When recipients of the Togo CCT were matched based on demographics to non-recipients, they were 66% more likely to access healthcare and less likely to use traditional medicine.51 While no other studies in this review reported on health service utilisation, some had secondary findings that suggested various effects. A study evaluating UNHCR UCTs in Jordan indicated that the transfers enabled some beneficiaries to partially cover the costs of treatment. While this may alleviate some of the costs of accessing healthcare, it was not decisive in accessing healthcare.52 This same issue was reported by a respondent in Palestine who stated ‘We still don’t have money for medicines, so we use the cash to pay off debts at the pharmacy’.29 This was further reflected in Bangladesh where, even though forecast-based UCTs were used the second most frequently on health expenses following food, it did not result in significant differences in the experience of illness compared with the comparison group (intervention 17.8% of 152; comparison 20.1% of 149, p value 0.60).40

### Enabling and constraining factors to cash transfer implementation

Only a few studies reported on enabling or constraining factors around cash transfer implementation. Enabling factors mentioned were lower costs for implementing agencies compared with other modalities (eg, vouchers),30 33 34 giving households the freedom of choice,30 enabling policy environments41 and beneficiaries’ preference for cash.34 Yet, several factors may constrain successful implementation including: weak health service infrastructure,31 reduced onsite access,35 dysfunctional markets or limited market choices,39 beneficiary decision-making (especially of caregivers),31 barriers to accessing cash programming by specific groups because of, for example, their gender or age,37 struggle to access cash due to travel limitations46 or a lack of concrete risk assessment tools that prevent meaningful consultation with community members into the programme design.37 Additionally, logistical challenges such as physical and institutional barriers to access, methods of disbursement and corruption or bribery were mentioned.37

## Discussion

This review presents evidence on the impact of cash on health in humanitarian settings. Most studies were on UCTs in human conflict or food crisis settings discussing their impact on nutrition, psychosocial and mental health or general health and well-being. While the purpose of several programmes was specifically aimed at improving food security and preventing malnutrition, few were specifically designed with as purpose to address other health outcomes.30–34 36 39 42–47 Only five studies reported on the utilisation of health services.35 48–51 Nineteen studies reported some positive impacts on various health outcomes and use of health services,29 31 33–38 40–45 47–51 11 reported no statistically significant impact on outcomes assessed30 35 36 38–40 44 46–49 and 4 reported potential negative impacts on health outcomes.36 37 42 45 While the majority of studies described nutrition and diet-related outcomes, the impact cash transfers had on these health outcomes varied. Although, most studies reported decreased frequency of acute malnutrition or related metrics (eg, dietary diversity) in cash groups compared with in-kind food, food voucher and/or no assistance, two studies found that cash beneficiaries consumed less calories than in-kind food beneficiaries.42 45 Furthermore, while most studies reported positive impacts on psychosocial and mental health, often related to alleviating the stress of financial burdens, two studies reported social exclusion of recipients and verbal abuse from non-beneficieries.36 37

Our findings are broadly in line with evidence from research on cash transfers’ impact on health outside of humanitarian settings. For example, a 2010 Cochrane review on CCTs in low and middle-income countries (LMICs) reported that despite methodological weaknesses, the evidence suggests that cash transfers may contribute to health benefits.10 Likewise, systematic reviews published 3 and 5 years later, respectively, on maternal, child and newborn health and CCTs suggested that cash transfers are effective in addressing child health determinants (eg, access to healthcare, morbidity risk),24 and that CCTs improved ANC visits, delivery at a health facility, skilled attendance at birth, reduction of low birth weight incidence and increased tetanus toxoid vaccination of mothers.24 When exploring the wider (grey) literature from a humanitarian perspective, a Humanitarian Policy Group 2012 report indicated that there is some evidence backing up the use of cash transfers to improve nutritional status in emergency settings if markets are functioning and quality food is available.53 Finally, a 2015 Cochrane review on cash transfers in humanitarian disasters in LMICs concluded that studies either reported improved outcomes or no statistically significant evidence of UCTs impacting health outcomes. However, they considered the body of evidence to be of very low quality with great uncertainty across all outcomes.22 The findings of this review are compatible with these previous reviews: there is evidence to suggest a positive relationship between cash transfers and health outcomes; however, there is a need for stronger additional high-quality evidence that can also be synthesised through a meta-analysis to determine the extent of this impact. Additionally, there may be unintended consequences from cash transfers that should be further explored.

The included studies were heterogeneous in their approaches to examine the effect of cash. Some studies investigated the impact of cash compared with food vouchers or in-kind food, while others used a comparison group that did not receive any assistance, tested the value/distribution method of cash transfers or performed pre–post implementation comparisons. Testing the implementation of cash transfers in one group against a group with no aid assistance also posed ethical challenges. Consequently, it is difficult to ascribe the extent or magnitude of the effect due to cash transfers versus other mechanisms. Studies comparing cash to vouchers, for example, often reported there was not a statistically significant difference in their impact. However, there is often a preference for cash transfers over other aid forms.13 14 With this in mind, policymakers and programme designers should consider the risks and benefits of these different approaches before implementing these interventions.

Included were limited in the time frames; studies were conducted over a few months (2 months) to a few years (18–24 months). Additionally, the length of cash transfers varied, from one-time disbursements to monthly allotments over a 7-month period. Going forward, it will be important to measure outcomes on a larger scale over a longer period of time, to fully understand whether cash can offer sustainable, long-lasting positive health impacts. Future research and aid provision could also consider the length of time covered by the cash transfers themselves and the effect of distributing the same monetary value over shorter versus longer time periods. The health outcomes examined across the studies were also fairly limited in scope, with only one study exploring the impact of cash on acute illnesses. This could be due to the requirement of laboratory-intensive or invasive measurements for some health outcomes in order to ascertain disease/health. Yet, cash transfers may also positively contribute to a range of other communicable (eg, COVID-19, Ebola) and non-communicable diseases. Several cash transfer programmes have been brought in place in the response to COVID-19 over 2020–2021 globally. This includes the expansion of the two largest existing cash transfer programmes in Colombia (Families in Action and Youth in Action) by lowering eligibility thresholds and including education and mental and psychosocial health targets.54 55 However, while several studies on cash transfers during COVID-19 were retrieved in our search strategy, the majority did not focus on the cash transfers’ impact on human health but rather on mitigating the economic impacts of the pandemic.56–59

Interestingly, the vast majority of cash transfers included were not specifically designed to cover health expenses. This has two important implications. First, there could be potential bias induced when cash is distributed for a specific purpose and communicated as such—resulting in recipients aiming to conform to what they were told the cash should be spent on (eg, food). Second, if the cash transfer value has not been designated to cover health expenses, households may likely trade-off and prioritise different expenses including health. As cash transfers become increasingly common, it will be important for implementers to collect, analyse and share the data on the effectiveness of their interventions in order to inform future programmes and evidence-based. Documenting best practices and considerations on safe and ethical implementation are important considerations. Therefore, it may be useful for future studies to adapt a similar framework or investigate a consistent group of core metrics in order to assuage some of the heterogeneity of this literature base. Additionally, future research and documentation of evidence could consider the roles of the different actors involved in the conceptualisation, development and delivery of cash transfers. It is important to consider how different implementers may have different motives and do not necessarily obey the same humanitarian principles, which can expose beneficiaries to different risks.60

The type of data collected and reported is also an important consideration; qualitative data may provide further detail and insight into the experiences and perspectives of recipients and implementers. The qualitative data synthesised here highlighted some of the possible unintended consequences or impacts such as social exclusion, community tension and verbal abuse.36 37 Humanitarian settings pose additional challenges that must be considered, in terms of collecting data and adapting the intervention in real time where appropriate.61 Innovative social science and anthropology methodologies in implementation research that emphasise the influence of context, and unique experiences of different populations and settings may prove useful in investigations of cash transfers in humanitarian settings.53 62 These approaches do this by providing not only a positivist or binary outcome about the effectiveness of programmes but also more experiential insight into what works and does not work.

One of the limitations of this study is that it focuses on cash and does not include vouchers, which have been increasingly used to improve accessibility to health facilities. Second, we focused on the direct impact of cash transfers on health outcomes and service utilisation without exploring wider social determinants indirectly affecting health, such poverty reduction, clean water and sanitation access and education. Third, the available evidence was limited, and studies often had significant limitations, complicating robust information synthesis and preventing the performance of additional analysis (eg, meta-analyses). Yet, opportunities for rigorous approaches in acute emergencies are limited due to inaccessibility and the short planning cycles of intervention design and implementation. Finally, we focused on the inclusion of peer-reviewed academic journals in order to limit potential biases and confounders, inaccuracies and incomplete information and to ensure the replicability of this review.63 Consequently, evidence reported in the grey literature (including NGO and government reports, theses and dissertations) that provide data not found in the peer-reviewed published academic literature may have been missed, resulting in a chance of publication bias as authors tend to publish studies with significant results. Despite this, the review has several strengths, including a detailed and updated comprehensive search strategy to gather available evidence, the synthesis of both quantitative and qualitative literature, unlimited date range of publications included and a broad definition of health outcomes allowing for a diverse examination of the impact of cash transfers on health. To our knowledge, this is the first systematic review on both conditional and unconditional cash transfers and health outcomes in humanitarian settings. The range of data presented in this review emphasises that the impact of cash transfers is not homogeneous across settings due to differences in exposure and nature of disasters, vulnerability, sensitivity and adaptive capacity of the population in a humanitarian setting. For example, both the transfer value and timeliness of distribution are components of cash transfers that influence their effectiveness.

The findings from this systematic review exhibit not only the potential impact of cash transfers on health outcomes and health service utilisation but also calls for future research. There is urgent need for high-quality quantitative and qualitative empirical evidence that is methodologically robust, investigates a range of humanitarian settings, and is conducted over longer time periods to better understand the long-term impacts of cash transfers on health and health service utilisation in humanitarian settings. Future research must investigate this area in further detail to better understand the specific variables that influence the effectiveness of cash transfers on health outcomes. For example, considering the types of crises (armed conflict vs epidemics) or health metrics (chronic vs infectious diseases). These lines of investigation could also provide insight to the impact of cash transfers on health outcomes beyond nutrition and mental/psychosocial health that were most examined in this review. Additionally, there is a need for further and clearer evidence on implementation factors that shape how cash transfers may function in a setting. For example, across health and international development interventions, it is highly encouraged and even expected to involve stakeholders directly at the beginning of a programme rather than to introduce an intervention from the outside or top-down approaches. Humanitarian or emergency settings may pose unique challenges when it comes to the timeliness and logistics of response and so future work may consider the role of building local capacity within cash transfer and other programme that can be leveraged in times of need. The findings of this review, as well as this call for further research, can have implications for both policy and practice by informing the development of evidence-based cash transfer programmes as they are implemented across humanitarian settings.

## Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information. Not applicable.

## Ethics statements

### Ethics approval

This research did not require an institutional review board approval as the data were collected from existing online databases and publicly available. This research did not involve any human subjects.

## Acknowledgments

The authors thank the peer-reviewers for their constructive feedback that enabled the improvement of the manuscript.

• ## Supplementary Data

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

## Footnotes

• Handling editor Seye Abimbola

• Contributors KRvD conceived the presented idea and developed the research protocol with support from SD, RJ, HCA, PK, JL, CM, YK, MYE and IK. KRvD, SD, RJ, HCA, PK, JL, CM, YK, MYE and IK collected, analysed, and synthesised the data. KB and HJ provided critical feedback and expertise on the protocol, analysis and write-up. All authors have made substantial, direct and intellectual contributions to the work and approved it for publication. KRvD is responsible for the overall content.

• Funding This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector. KRvD received funding from the Gates Cambridge Trust (OP114) for her PhD studies and received funding for publication of this article from the Gates Foundation.

• Competing interests None declared.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.