Application of the Three Delays Framework to Assess Barriers to Surgical Care in an East African Refugee Camp: A Cross-sectional Survey Among 1,135 Adults

Abstract

Introduction

The three delays framework assesses barriers associated with seeking (delay 1), reaching (delay 2), and receiving (delay 3) healthcare in global health. This study applies this model to identify factors contributing to delays in surgical care among adult refugees in Tanzania.

Methods

A validated surgical needs survey was administered using random cluster sampling to adults aged 19 and over living in Tanzania’s Nyarugusu refugee camp. Participants with potentially surgically amenable issues were categorized by whether they sought, reached, and/or received care promptly for each concern. Multivariate logistic regressions identified factors that were significantly associated with the three delays.

Results

There were 1,135 participants with 2,463 potentially surgically amenable concerns. Of these, 689 adults (60.1%) experienced a delay for at least one problem, with care-seeking delays being most common (n=477, 42.2%). Illiteracy, no prior use of the camp health center, traditional healer usage, and having a burn were associated with delays in seeking care. Older age and positive self-perception of health status were associated with delays in reaching care, while lack of recent illness and Burundian nationality were correlated with delays in both seeking and reaching care. Female sex and Congolese nationality were associated with delays only in receiving care, while having a wound was correlated with delays in reaching and receiving care.

Conclusion

Adults in Tanzania’s Nyarugusu refugee camp experience unmet surgical needs attributable to multifactorial barriers along the surgical care continuum. Raising patient awareness of surgically treatable conditions and increasing health services accessibility may improve health outcomes for refugees in humanitarian settings.

Introduction

An estimated 16.9 million people die annually from surgically amenable conditions, which is higher than the pooled mortality from HIV/AIDS, tuberculosis, and malaria.^1,2 Low-to-middle-income countries (LMICs) experience the greatest burden of untreated surgical disease, with the prevalence of ranging from 11% to 46%, depending on the country and study population.^3–5 Various barriers, including limited surgical capacity, distance to health facilities, and financial constraints contribute to these disparities.^5,6 However, there has been limited attention to characterizing barriers to surgical care among forcibly displaced populations, including the 18 million refugees in sub-Saharan Africa, who face a greater burden of disease than non-refugee populations.^7,8

The three delays framework has been used to characterize barriers to seeking (delay 1), reaching (delay 2), and receiving (delay 3) care in various arenas of global health, including maternal mortality and surgery.^9,10 This model can be useful in more accurately identifying barriers to care and in guiding interventions and policy. Concepcion et al. (2020) recently applied the three delays framework to characterize barriers related to geographic distance from health facilities, inadequate healthcare infrastructure, and utilization of traditional healer services among children in Somaliland.¹¹ However, in refugee camp settings, delays in reaching care may be less prominent, while delays in seeking and receiving care are very commonplace. As such, the three delays model provides a useful tool to assess specific challenges related to surgical care for adults in humanitarian settings that have not been characterized previously.

A previous analysis conducted by our team found that 46.3% of refugees in Nyarugusu shared at least one health concern that may be amenable to surgical intervention and over 60% reported an ongoing health problem that may be surgically correctable but had not been addressed.⁴ This study builds on prior work by identifying barriers and delays in accessing surgical care for adult refugees in Tanzania, including those without prior contact with formal healthcare. Context-specific findings — highlighting challenges such as limited prior access and cultural perceptions of biomedical health services — can inform policies and interventions aimed at improving surgical access in refugee camps and humanitarian settings.

Methods

Study setting

Established in 1996, Nyarugusu refugee camp is located in the Kigoma region of western Tanzania and is home to approximately 133,000 refugees, mostly from the Democratic Republic of Congo (DRC) and Burundi, as of 2024.¹² At the time of this study, the camp was divided into 14 zones, which were further subdivided into villages and clusters. Over 10,000 operations were performed in the camp’s operating theatre(s) from 2000–2020, with cesarean sections and herniorrhaphies and being the most common procedures.¹³

Data collection

This analysis is based on data from a parent study conducted from August 4, 2021 to September 10, 2021 that assessed the burden of surgical disease among refugees living in Nyarugusu. The parent study surveyed 3,574 refugees about their experiences navigating illness and health resources in the camp. A target sample size of 3,519 participants was determined for the parent study through a standard prevalence-based calculation. The following formula was used: n = Z² p(1 − p) / L² with a Z-score of 1.96 corresponding to a 95% confidence level, the expected prevalence of surgical conditions (p), and the precision required around that estimate (L). Based on prior literature, we assumed an 11% prevalence and specified a 1.5% margin of error. To account for cluster sampling, the calculated value was doubled using a design effect of two, and an additional adjustment of 5% was included to accommodate anticipated non-response, while maintaining adequate statistical power. For additional details this sample size calculation, please refer to the parent study.⁴ Furthermore, 132 clusters were randomly sampled among 1,472 clusters based on the administrative divisions of zones, villages, clusters, and households. Within each cluster, every household was approached, and two individuals from each household were randomly selected to participate.

The data collection process was carried out by refugee community healthcare workers (CHWs). The Surgeons Overseas Assessment of Surgical Needs (SOSAS) tool was used to evaluate surgical needs and barriers, with modifications to include questions about referral patterns.¹⁴ This tool is a validated survey instrument previously used in various low- and middle-income countries (LMICs) and refugee settings.¹⁵ The CHWs used a translated (Kiswahili) version of the SOSAS tool to perform verbal head-to-toe exams and to identify surgical problems across anatomical regions (e.g., head/neck, chest, abdomen, extremities). In clusters where Kirundi or other indigenous languages was the main language, CHWs with command of both Kiswahili and the other language were used. Participants were also asked about potentially surgically amenable issues such as wounds, masses, and deformities, and whether they had prior surgeries, such as caesarean sections.

After excluding records with technical issues (e.g., 99% missing data) or consents that were missing or not provided by the patient (n=36), 3,574 usable records remained from an initial 3,611. For this study, additional exclusions were made for individuals 18 years old and under, as children were seen as a distinct population with unique needs and barriers. Patients with conditions or concerns deemed not amenable to surgery were also excluded. Data collection was conducted offline using REDCap Mobile and uploaded to the secure Johns Hopkins University REDCap data server. The data was then screened for quality by a locally hired research manager. Full details on sampling methods and data collection are available in the parent study.⁴

This research was based on a subset of 1,135 responses from the parent dataset based on the aforementioned inclusion and exclusion criteria. Respondents or their close relatives (e.g., adult children) self-reported conditions potentially amenable to surgical intervention. Surgical conditions were defined according to the Lancet’s criteria as “any disease, illness, or injury in which surgical care can potentially improve the outcome”.^1,16 In this study, healthcare-seeking was defined as utilizing biomedical care offered at physician-based or formal health centers. In contrast, seeking non-biomedical care, such as treatment by traditional healers or other alternative modalities along the surgical care continuum, was analyzed as a separate variable and categorized as a delay in seeking biomedical care. Respondents who sought biomedical care at clinics or hospitals were asked to describe the care received for their condition and then community health workers classified the care into one of four categories: (1) no healthcare or surgery, (2) healthcare received (e.g., medications) without surgery, (3) minor surgical procedures, or (4) major surgical procedures.

Applying the Three Delay Model

The first delay (D1)—delays in seeking care—was defined as not pursuing care for any reason, including consultation of a traditional healer, perceiving no need for care, experiencing fear, facing financial or knowledge barriers, lacking access to health services, or facing time constraints. The second delay (D2)—delays in reaching care—was defined as taking more than an hour to travel to an in-camp health facility and/or waiting over an hour for transportation. A one-hour cutoff was used based on contextual and logistic considerations within the camp. For example, Nyarugusu’s main hospital is centrally-located in a relatively compact camp and accessible by foot for an average adult without physical impairments. Therefore, needing more than one hour to reach care would indicate significant geographic, physical, or logistical barriers, such as mobility challenges, lack of transportation, or environmental obstacles. The third delay (D3)—delays in receiving care—was defined as experiencing a wait time of greater than six hours after arriving at an in-camp health facility before receiving treatment. Though prolonged wait times may be expected given the camp’s resource limitations, the cut-offs for D2 and D3 are not intended to characterize clinically significant delays but rather to reflect a perceived potential barrier from the patient or caregiver’s perspective, which may influence future decisions to seek care. Furthermore, it is important to note that these thresholds are context-specific and may be appropriate for Nyarugusu refugee camp but should not be generalized to other humanitarian settings or be seen as universal, clinically meaningful benchmarks.

Data analysis and ethical approval

All statistical analyses were conducted using R version 3.0.1 (R Foundation for Statistical Computing, Vienna, Austria). Descriptive analyses utilized t-tests and Chi-square tests, with continuous variables reported as means ± standard deviation and categorical variables as counts and percentages. A p-value < 0.05 was considered statistically significant. Multivariable logistic regression was performed to identify associations between covariates and delays in seeking, reaching, and/or receiving surgical care. Odds ratios with 95% confidence intervals were reported, with standard errors adjusted for the clusters in the analysis. Covariates included in the regression models were age, sex, nationality, literacy status, current health status, recent primary health center (PHC) utilization, condition pathology, and illness within the past year. Zone, education, religion, occupation, and marital status were excluded due to collinearity or insufficient responses, which caused inflated standard errors. Additionally, because participants who had not previously used a PHC did not have transportation times available to classify whether they experienced delays in reaching or receiving care, prior use of a PHC was not included in these multivariable models. Participants were analyzed at the individual level, however as each participant could endorse multiple issues, each medical concern was treated as a separate entry to fully capture the range and prevalence of all presenting conditions. For this reason, a single participant could be associated with multiple delays—either for different conditions or multiple delays related to the same condition. This condition-level approach was chosen to capture pathology-specific delays and better reflect the camp’s spectrum of surgically amenable disease. Patient-level aggregation was considered but would have made it more difficult to identify these pathology-specific associations relevant to surgical access.

The study received approval from the Johns Hopkins Medicine Institutional Review Board (IRB00258009) and research clearance from the Tanzanian Commission on Science and Technology (2020–391-NA-2011–143). Permission to access the refugee camp was granted by the Tanzanian Ministry of Home Affairs. Informed consent or assent was obtained from all participants or their household proxies.

Results

Demographics

Among 1,135 participants, 2,463 different concerns potentially amenable to surgical intervention were self-identified by respondents. The participants’ mean age was 38.8 years (±15.7), with the largest age group being 19–39 years (n = 659, 58.1%) (Table 1). Most participants were female (n = 768, 67.7%), and a slight majority were Congolese (n = 597, 52.7%), followed by Burundian (n = 531, 46.9%). Nearly one-third of respondents were not literate (n = 347, 30.6%), and the most common level of education completed was secondary school (n = 413, 36.4%). Most participants were also Christian (n = 1,045, 92.1%), single (n = 724, 63.8%), unemployed (n = 513, 45.2%) or worked as farmers (n = 285, 25.1%).

Table 1.

Demographic profile of the study population

	Characteristic	Total (n=1135) (%)
Age, mean ± SD		38.8 ± 15.7
Age Category	19–39	659 (58.1)
	40–59	310 (27.3)
	60 and over	166 (14.6)
Zone	1	168 (14.9)
	2	84 (7.4)
	3*	18 (1.6)
	4	80 (7.1)
	5	64 (5.7)
	6	154 (13.6)
	7	31 (2.7)
	8	206 (18.2)
	9	187 (16.5)
	10	36 (3.2)
	11	74 (6.5)
	12	16 (1.4)
	13	4 (0.4)
	14	8 (0.7)
	Unknown	5 (0.4)
Sex	Female	768 (67.7)
	Male	367 (32.3)
Nationality	Burundi	531 (46.9)
	Democratic Republic of the Congo	597 (52.7)
	Unknown	7 (0.6)
Religion	Christian	1045 (92.1)
	Muslim	60 (5.3)
	Other	28 (2.5)
	Unknown	2 (0.1)
Education	Uneducated	315 (27.8)
	Primary	379 (33.4)
	Secondary	413 (36.4)
	Higher Education	27 (2.4)
	Unknown	1 (0.1)
Literacy Status	Literate	787 (69.3)
	Illiterate	347 (30.6)
	Unknown	1 (0.1)
Marital Status	Single/Unmarried	144 (12.7)
	Married	724 (63.8)
	Divorce	101 (8.9)
	Other	161 (14.2)
	Unknown	5 (0.4)
Occupation	Small businesses	90 (7.9)
	Self-employment	79 (7.0)
	Housewife	98 (8.6)
	Farmer	285 (25.1)
	Unemployed	513 (45.2)
	Domestic Helper	8 (0.7)
	Other	62 (5.5)
	Unknown	0 (0.0)

SD Standard deviation

^*Nyarugusu’s main health center is located in Zone 3

Health status and surgical pathology

Most participants self-reported generally good health (n = 724, 63.8%), but also endorsed having an illness (n = 846, 74.5%) within the past year (Table 2). A high proportion reported using the camp’s formal health services (i.e., hospital or clinics) within the past year (n = 1,084, 95.5%), but many also sought care from traditional healers (n = 193, 17.0%). A significant proportion reported only one health concern (n = 492, 43.3%), but 643 (56.7%) had multiple issues.

Table 2.

Health status, surgical problems, and anatomical locations of the study population

	Characteristic	Total (n=1135 (%))
Self-perceived good health status	No	398 (35.1)
	Yes	724 (63.8)
	Unknown	13 (1.1)
Prior use of Primary Health Center within past year	No	49 (4.3)
	Yes	1084 (95.5)
	Unknown	2 (0.2)
Illness within the past year	No	283 (25.0)
	Yes	846 (74.5)
	Unknown	6 (0.5)
Previously sought care from traditional healer	No	942 (83.0)
	Yes	193 (17.0)
Pathology of Issue	Wounds only	217 (19.1)
	Congenital Deformities only	175 (15.4)
	Burns only	74 (6.5)
	Swelling/tumor only	202 (17.8)
	Other	123 (10.8)
	Multiple	344 (30.3)
Number of Concerns	1	492 (43.3)
	2	320 (28.2)
	3	144 (12.7)
	4	76 (6.7)
	5 or more	103 (9.1)
Ongoing Concern
	Yes	747 (65.8)
	No	388 (34.2)
Anatomical region affected *
	Head/Face/Neck	650 (57.3)
	Groin	283 (24.9)
	Breast/Chest	241 (21.2)
	Abdominal	287 (25.3)
	Extremity	427 (37.7)
	Back	242 (21.3)

^*This variable represents the number and proportion of people who had at least one problem within the specified anatomical region out of a total of 1,135 participants. If a respondent had multiple problems within the same region, then the region was counted once.

Regarding self-identified concerns that may have been amenable to surgical intervention, 217 patients (19.1%) had wounds only, 175 (15.4%) had only acquired deformities from birth or aging, and the rest had only burns (n = 74, 6.5%), or swelling/tumors only such as goiter (n = 202, 17.8%). Those with only unspecified pathologies were listed as “Other” (n=123, 10.8%). Many participants self-reported multiple problems with a mix of pathologies (n = 344, 30.3%), including those that were specified and unspecified. Of the 1,135 patients, 747 (65.8%) reported an ongoing or untreated problem. A slight majority sought care for all concerns (n = 652, 57.4%), while 230 (20.3%) deferred care for at least one problem, and 253 (22.5%) did not seek care at all. Among all participants, concerns within the face/head/neck region were most common (n = 650, 57.3%), followed by those in the extremities (n = 427, 37.7%).

Three Delays Model

Among the 1,135 respondents who self-reported health concern(s) that may have been surgically intervenable, 689 adults (60.1%) experienced at least one delay along the surgical care continuum. Specifically, 477 adults (42.2%) delayed seeking care for at least one concern (D1), 169 respondents (14.9%) experienced a delay in reaching care (D2), and 57 (5.0%) experienced both D1 and D2. Delays in receiving care (D3) were reported by 233 respondents (20.5%), among whom 78 (6.9%) also delayed seeking care and 72 (6.3%) also experienced delays in reaching care. A small minority of participants (n = 17, 1.5%) reported delays across the entire surgical care continuum (Figure 1). Among 643 participants with multiple health concerns that may have been amenable to surgical intervention, 207 experienced different types of delays (e.g., D1 for one problem and D3 for another). The most cited reasons for delayed biomedical care-seeking included consultation with a traditional healer instead (n = 88, 18.4%), a lack of perceived need for care (n = 30, 6.2%), and a perception that healthcare services were unavailable (n = 30, 6.2%) (Table 3). Ultimately, 223 (19.6%) participants received either minor or major surgery for at least one of their issues, but among those who received surgery, delays were nonetheless common: 44 had at least one delay in seeking (n = 12), reaching (n = 10), and/or receiving care (n = 27).

Figure 1.

Prevalence of delays for impacting surgical care (n = 1,135). Note: D1, delay in seeking care; D2, delay in reaching care; D3, delay in receiving care. The sizes of the circles are proportional to the prevalence of the delay. Overlaps in circles represent patients who have a combination of those delays.

Table 3.

Cited factors contributing to delays in seeking care

	Experienced D1 (delays in seeking care, n = 477(%))*
No time	2 (4.2)
Fear/lack of trust of PHCs	10 (21.0)
No perceived need for formal care	30 (6.2)
Sought care from traditional healer	88 (18.4)
No money for health care**	1 (2.1)
No money for transportation	1 (2.1)
Do not know where to find treatment	2 (4.2)
Health care unavailable	30 (6.2)
Not otherwise specified	423 (88.7)

PHC Primary health center

^*This variable was categorized by the number of participants who cited the listed reason for at least one specific concern. If the same concern applied to multiple problems, then it was counted once. The proportion was then calculated by dividing the counts for each individual reason with the total number of people with a delay in seeking care.

^**Healthcare services in the camp are provided free of charge.

Regarding delays in reaching (D2) and receiving care (D3), a small proportion of respondents reported a travel time of one to two hours to the health center (n = 143, 12.6%) while 56 (4.9%) waited more than an hour for transportation. Pooling those who waited at least an hour for transportation and/or travel time to the hospital, 169 (14.9%) participants experienced a delay in reaching care. At the health center, the most reported waiting period was between three and five hours (n = 494, 43.5%). However, 233 patients (20.5%) waited more than six hours and were therefore considered to have experienced a delay in receiving care (Table 4). Additionally, 51 (4.5%) adults had no reported waiting or transportation times available, as they had not previously used a primary health center (PHC).

Table 4.

Travel time and waiting time at the hospital

	Characteristic Total	(n=1135) (%)
Time to travel to in-camp health center
	Less than 1 hours	935 (86.3)
	1 −2 hours	143 (13.2)
	3–5 hours	4 (0.4)
	6–12 hours	1 (0.1)
	Did not go to a health center	49 (4.3)
Time waiting for transportation
	Less than 1 hours	1028 (90.6)
	1 −2 hours	46 (4.2)
	3–5 hours	8 (0.7)
	6–12 hours	2 (0.2)
	Did not go to a health center	51 (4.5)
Time waiting to be seen at the health center
	Less than 1 hours	128 (11.8)
	1 −2 hours	229 (21.1)
	3–5 hours	494 (45.6)
	6–12 hours	178 (16.4)
	More than 12 hours	55 (5.1)
	Did not go to the health center	51 (4.5)

Factors Associated with Delays

Factors significantly associated with delays in seeking surgical care (D1) included nationality, with Congolese adults being more likely to seek care than Burundian participants (aOR: 0.72; 95% CI: 0.55–0.95). On the other hand, being literate (aOR: 0.68; 95% CI: 0.50–0.93), having experienced an illness in the past year (aOR: 0.65; 95% CI: 0.47–0.90), using a PHC previously (aOR: 0.43; 95% CI: 0.22–0.81), and deferring initial consultation with traditional healers (aOR: 0.49; 95% CI: 0.34–0.70) predicted positive health-seeking behavior. Regarding the pathology of concerns, participants with burns (aOR: 6.34; 95% CI: 3.35–12.6) and those with multiple types of concerns (aOR: 1.84; 95% CI: 1.25–2.72) were more likely to delay seeking care (D1) compared to those with wounds (reference group).

For delays in reaching care (D2), the most strongly associated factor was pathology. Participants with burns (aOR: 0.07; 95% CI: 0.00–0.32), congenital or acquired deformities (aOR: 0.34; 95% CI: 0.19–0.62), and multiple types of conditions (aOR: 0.28; 95% CI: 0.16–0.46) were more likely to reach care quickly compared to those with only wounds. However, delays in reaching care were more common among participants aged 60 and older (aOR: 1.8; 95% CI: 1.02–3.14), Burundian participants (aOR: 2.0; 95% CI: 1.3–3.13), those who perceived themselves as healthy (aOR: 1.47; 95% CI: 1.01–2.15), and those who had illnesses in the past year (aOR: 2.48; 95% CI: 1.50–4.32).

Regarding delays in receiving care (D3) at health facilities, Congolese participants (aOR: 2.13; 95% CI: 1.52–3.00) and females (aOR: 2.04; 95% CI: 1.39–3.00) were more likely to experience delays. Pathology also played a role: patients presenting with burns (aOR: 0.25; 95% CI: 0.07–0.67) and swellings/masses (aOR: 0.43; 95% CI: 0.24–0.75) experienced fewer delays in receiving care than patients with wounds.

Discussion

This study highlights a high degree of delays in accessing care among adult refugees living in Nyarugusu with self-reported health concerns that may be amenable to surgical intervention, most notably with delays in seeking and receiving care. This deeper understanding of the primary associated factors or contributors leading to delayed surgical care in refugee camps can help inform future research and policy changes in this humanitarian setting and others.

Delays in seeking care

Nearly half of the adults categorized into delay 1 did not seek care for any of their health concerns. Thrush and Hayder (2014) and de Siqueira Filha et al. (2022) suggested that adults may deprioritize their health due to work, caregiving responsibilities, or economic pressures, but few in our cohort reported not having time (n=2) nor money to reach or receive care (n=2).^17–18 The low proportion of patients citing money as a concern may be because all healthcare services within the camp are free of charge. Thirty patients reported no perceived need for care, which may be due to a lack of awareness and belief that these issues are manageable or not urgent.¹⁹ However, 423 (88.7%) participants did not select one of the eight possible reasons for not seeking care for at least one of their problems. This high proportion of unspecified reasons limits interpretability and creates uncertainty about other potential factors that may be influencing refugees’ healthcare-seeking decisions but were not captured by the survey. One hypothesis is that forcibly displaced persons living in extremely resource-limited settings may worry that they will not receive definitive care given severe limitations on trained personnel and health care resources available in camp. The referral system to hospitals outside of the camp is also severely under-resourced; many patients wait months or years before receiving a referral, the outcome of which may or may not be therapeutic.²⁰ Other unspecified reasons could include perceived discrimination, language/cultural differences, and stigma surrounding their condition (such as genital deformities).^21–23

Refugees who consulted traditional healers were more likely to delay seeking care at a PHC. We could not determine whether traditional care was sought before or after a patient decided to seek care at a PHC or was sought simultaneously. However, in other countries, traditional healers are often the first-line providers for various concerns.^24–26 This preference for alternative forms of healing is often due to a perceived lack of available biomedical services, familiarity and trust in traditional practices, lower costs, and greater alignment with cultural values, and may apply to Nyarugusu as well.^26–28 As such, collaborative partnerships between physicians, community health workers, and traditional healers could improve the timeliness of referrals for concerns requiring biomedical treatment while building trust with community members.²⁹ The protective value of literacy and prior engagement with formal healthcare services also highlight the importance of supporting health education and outreach programs with community health workers. These programs increase patient knowledge and confidence regarding when to seek care and can facilitate initial engagement with biomedical services.^30–32 Furthermore, participants with burns were more likely to delay seeking care, relative to wounds, reflecting potential severity bias rather than true access barriers. This may because some wounds are often highly visible and painful but mild burns may be initially managed at home or through traditional remedies while biomedical care is sought for more severe ones.

Delays in reaching care

Delays in reaching care were most closely associated with older age, positive self-perception of health, and wound pathology. Though Nyarugusu’s primary health center is accessible by foot for most patients, older adults may experience high rates of poverty contributing to limited ability to pay for transportation and functional impairments from chronic conditions.^33–34 Those who perceived themselves as healthy may also underestimate of the severity of their conditions.³⁵ Additionally, patients who lived in more geographically distant zones may also have had greater difficulty reaching the camp during the rainy season. Our data, however, did not account for the variance in seasonal climate and in terrain across the zones. Though travel distance was not theoretically prohibitive overall, delays in reaching care in other contexts can contribute to increased readmissions, lower survival rates, worse follow-up adherence, etc.^36–37 Individuals with burns and multiple types of pathologies were more likely to reach care promptly relative to wounds, reflecting potentially, as mentioned above, the perceived urgency or visible severity of some acute burns, prompting faster movement through the biomedical healthcare system. Though Botman et al. found that 80% of burn victims may reach treatment promptly in Tanzania, they reported that only 50% reach a facility providing surgical burn care within 24 hours. While we lacked data on burn referrals and severity, lack of trust and surgical capacity were cited as key issues driving delays in addition to limited burn care infrastructure in Tanzania.³⁸ These findings suggest that targeted interventions such as free community-based transportation services and health education programs for the elderly may reduce delays in reaching care.

Delays in receiving care

Perceived delays in receiving care were most strongly associated with female sex, Congolese nationality, and wound pathology. These findings were surprising as camp health services in the camp are free-of-charge and open to patients of every age, religion, gender, or nationality. However, there may be cultural challenges at play. Regarding the sex-based discrepancy, women living in patriarchal societies may experience delays due to provider attitudes that dismiss female patients’ concerns, a lack of communication with hospital staff, trauma from sexual violence, or sex-based cultural norms.^39–42 For nationality-based differences, previous research from our group found statistically significant differences in education between Congolese and Burundian patients but no other differences in demographics.²⁰ However, Congolese refugees have settled in Nyarugusu for longer, and they may have accessed care when the camp's health system was serving a smaller camp population and less strained. They may consequently be perceived differently by healthcare staff compared to more newly arrived Burundian patients who may have more acute or previously unexamined needs. Language services are available, but given the wider range of indigenous languages found in the DRC, variations in cultural and linguistic competency among healthcare providers may also have contributed to perceived delays in receiving care.³⁹ Ultimately, our dataset does not capture these perceptions nor does the survey assess patient experience, provider behavior, or perceived discrimination directly. Similar disparities have been reported among different ethnic or national groups in other refugee or LMIC populations.^43–45 However, our interpretation should be viewed as plausible sociocultural or systemic hypotheses that require further qualitative study or direct testing rather than explanatory evidence of discrimination. Additional exploration of length of residence in the refugee camp, language issues, or administrative challenges with healthcare staff capacity and shortages is also warranted to further explain these differences.

These delays can lead to depression and feelings of isolation which may discourage future health-seeking behavior.^42,46 To address this, cultural competency training for hospital staff and female-led patient companion programs align with strategies to enhance patient-provider trust and to improve navigation within health systems.⁴⁷ Interestingly, patient with burns and swellings were less likely to experience delays in receiving care compared to wounds, which may also be depend on perceived severity and urgency as mentioned earlier. Creating specific wound care clinics or mobile wound care units that can perform debridement and assess for infection can help mitigate delays and reduce patient volume at the primary healthcare facility.⁴⁸

Limitations

This study had several limitations. Surveys were conducted by community health workers (CHWs) trained to identify possible surgical issues based on respondent-reported symptoms (e.g., swelling, burns, wounds, disabilities, or prior surgeries) rather than through comprehensive physical examinations by licensed medical professionals. This reliance on oral assessments may have resulted in over- or underestimations of the prevalence of surgical conditions. Future research could address this limitation by equipping CHWs with mobile technology to consult surgeons, reducing delays in care by bridging knowledge gaps, and addressing perceived needs within families. Sampling challenges further impacted the study as households unavailable for sampling due to repatriation were excluded, which may have led to underestimations of the surgical burden or the number of patients who did not have delays and were healthy enough to eventually relocate. The cross-sectional design excluded the possibility of following patients longitudinally to better characterize their disease experience and journey in the healthcare system. Adults with severe conditions who may have died before the survey period could have also contributed to an underestimation of the true surgical burden. Additionally, the reliance on self-reported data introduced recall bias, potentially underestimating or overestimating the prevalence of surgical problems and reasons for delayed care. We also did not ask how long patients had resided in the camp, which may influence health-seeking behaviors and contribute to the delays we saw herein.

Additionally, the time thresholds used to define delays in reaching and receiving care were context-specific and based on logistical realities within the camp rather than clinical meaningful cut-offs. These thresholds reflect perceived barriers to care and may not translate to other refugee or non-refugee settings, limiting external validity. Many respondents did not specify reasons for delays, potentially because of the limited number of reasons that respondents could select from or because surveyed household members were hesitant to cite a reason or lacked detailed knowledge of the issues—particularly if affected individuals were at a hospital or if the problem had occurred long ago. This limits the ability to fully characterize barriers to surgical care. “Wounds” were also used as the reference group for pathology-based comparisons, but they are heterogenous in nature, with varying acuity, visibility, and perceived urgency, ranging from minor lacerations to severe avulsions. Future studies would benefit from more granular classification of the wounds and pathology severity to better distinguish between delays driven by acuity or access. We also treated multiple conditions from the same participant as separate observations, which introduces potential non-independence and may inflate precision despite cluster-adjusted standard errors. Lastly, we did not record distance to health facilities, which prevented a geospatial analysis of households and health centers within the camp. Future research, examining geographic factors such as physical distance, terrain, or climate variations, can elucidate their impact on health-seeking behavior and accessibility as well as their contribution to delays in seeking and reaching care.

Conclusion

This study highlights an urgent need to address various barriers along the surgical care continuum, which impact over half of the sampled adults in Tanzania’s Nyarugusu refugee camp. Future research should explore longitudinal trends in surgical access and examine the role of traditional healers and referral processes in delays. More work is also needed to identify and evaluate potential interventions targeting specific contributors to these delays, which can hopefully improve timely access to surgical services and reduce this high burden of unmet surgical needs. Applying the three delays framework to other humanitarian settings can further refine strategies to improve surgical outcomes among displaced populations worldwide.

Table 5.

Multivariable analysis of factors associated with delays in seeking, reaching, and receiving surgical care among adults in Nyarugusu refugee camp (Aug.-Sept. 2021)

	DELAY 1		DELAY 2		DELAY 3
Characteristic	Odds Ratio	95% CI	Odds Ratio	95% CI	Odds Ratio	95% CI
Age Group
19–39	—	—	—	—	—	—
40–59	1.09	0.79, 1.50	1.27	0.82, 1.97	1.04	0.71, 1.52
60 and over	1.13	0.75, 1.71	1.8**	1.02, 3.14	0.92	0.55, 1.54
Gender
Female	—	—	—	—	—	—
Male	1.14	0.85, 1.53	1.14	0.85, 1.53	0.49**	0.33, 0.72
Nationality
Burundi	—	—	—	—	—	—
DRC	0.72**	0.55, 0.95	0.5**	0.32, 0.77	2.13**	1.52, 3.00
Self-perceived health status
No	—	—	—	—	—	—
Yes	1.06	0.79, 1.42	1.47**	1.01, 2.15	1.24	0.86, 1.78
Literacy Status
No	—	—	—	—	—	—
Yes	0.68**	0.50, 0.93	1.03	0.69, 1.55	1.13	0.77, 1.66
Illness in the past year
No	—	—	—	—	—	—
Yes	0.65**	0.47, 0.90	2.48**	1.50, 4.32	1.2	0.81, 1.79
Pathology of Wound
Wounds only	—	—	—	—	—	—
Burns only	6.34**	3.35, 12.6	0.07**	0.00, 0.32	0.25**	0.07, 0.67
Deformity only	1.39	0.90, 2.16	0.34**	0.19, 0.62	1.2	0.73, 1.96
Swelling/Growths	1.4	0.91, 2.14	0.61	0.36, 1.02	0.43**	0.24, 0.75
Multiple	1.84**	1.25, 2.72	0.28**	0.16, 0.46	1.01	0.65, 1.57
Other	0.79	0.47, 1.32	0.74	0.41, 1.32	0.59	0.32, 1.06
Sought Traditional Health
No	—	—	—	—	—	—
Yes	2.03**	1.42, 2.91	1.28	0.78, 2.06	1.22	0.81, 1.82
Religion
Christian	—	—	—	—	—	—
Muslim	1.16	0.64, 2.10	0.6	0.14, 1.83	0.47	0.11, 1.43
other	1.14	0.49, 2.59	0.84	0.30, 1.94	0.52	0.21, 1.14
PHC Use
No	—	—	—	—	—	—
Yes	0.43**	0.22, 0.81	—	—	—	—

CI Confidence Interval; DRC Democratic Republic of the Congo; PHC Primary Health Center

^*Odds Ratios quantify the likelihood of experiencing a delay in seeking, reaching, or receiving care in one group relative to another, with values greater than 1 indicating higher odds of delay, values less than 1 indicating lower odds, and a value of 1 suggesting no difference

^**Indicates an odds ratio that is statistically significant with a p-value <0.05

Data availability:

Data used and analyzed in the current study are not publicly available due to privacy and personally identifiable health information. De-identified, aggregate data are potentially available from the corresponding author upon request.