We conducted a multicenter, hospital-based, cross-sectional study over a 2-week period in January 2026. The study was carried out at the only two public hemodialysis (HD) facilities in Douala, the economic capital of Cameroon.

The HD unit of the Douala General Hospital (DGH): Established in 1990, it is the oldest and largest center in the country. The unit operates from Monday to Sunday with four shifts per day, except on Sundays, which run two shifts. An on-site blood bank is available within the institution. The HD unit of the Laquintinie Hospital of Douala (LHD): Operational since 2023, this center runs four shifts per day from Monday to Sunday. An institutional blood bank is also available on site.

All consenting adult patients (aged > 18 years) undergoing maintenance hemodialysis for more than 3 months were eligible for inclusion. We excluded patients with cognitive impairment, those unable to understand French or English, and “transient” patients who were regularly dialyzed at another HD facility outside the study sites.

During the study period, a total population of N = 414 patients on maintenance hemodialysis was registered across both centers (283 at DGH and 131 at LHD). To estimate the minimum sample size, we first applied Cochran’s formula for an infinite population.

where:

Z = 1.96, corresponding to the 95% confidence interval. e = 0.05 represents the margin of error. p = 0.82 (82%) is the expected prevalence of blood transfusion in HD patients in Cameroon, based on previous data reported by Fouda et al. [8].

This yielded an initial sample size (n₀) of 226.8. We then applied the finite population correction formula:

This determined a minimum required sample size of n = 147 patients.

Patients were approached in the waiting room before or after their dialysis session. We used consecutive sampling. A total of 225 patients were screened during the study period (DGH n = 138 and LHD n = 87). Of these, 56 were excluded: 11 were under 18 years old, 7 presented with cognitive impairment, 5 were transient patients, and 34 declined to participate. Ultimately, 168 patients (DGH n = 108 and LHD n = 60) were included in the final analysis (Figure 1).

Figure 1. Flowchart of participants.

Data were collected through face-to-face interviews conducted immediately before or after the dialysis sessions using a pre-tested, standardized questionnaire. Medical records were reviewed concurrently to cross-check and complete clinical, logistical, and biological variables. The collected variables were categorized as follows:

Sociodemographic Characteristics: Sex, age, area of residence, marital status, education level, and socioeconomic status (including monthly income and household wealth). The household wealth measurement was adapted from the World Bank Household Wealth Index, which assesses economic status based on a list of goods and services (housing quality, housing and vehicle ownership, water and electricity supply, number of persons per bedroom). A score (Si) was attributed to each asset, and the wealth index score was calculated using the formula

Where:

ΣSi = sum of the individual scores of each item.p/r = number of persons per bedroom.

Percentile distribution was subsequently used to categorize participants into three socioeconomic strata.

Participants belonging to the lowest two quintiles were classified as poor households.

Participants belonging to the highest quintiles correspond to wealthy households. Participants belonging to the 3rd and 4th quintiles correspond to intermediate households. Hemodialysis and baseline clinical characteristics: Blood group (ABO and Rh typing), hemodialysis vintage, type of vascular access (arteriovenous fistula or central venous catheter), and frequency of dialysis sessions per week. Anemia Management: The most recent pre-dialysis hemoglobin level, history of anemia therapies, utilization of ESAs along with the specific type used and compliance with prescriptions, and history of blood transfusions (number of blood units received during the past 3 - 6 months, average monthly requirements, type of blood product (whole blood vs. packed red blood cells)). Logistical and financial burden of blood transfusion: direct medical costs (estimated cost per blood unit, crossmatching fees), logistical parameters (use of an intermediary, type of replacement donor), and indirect costs (transportation expenses, estimated loss of earnings). The estimated loss of earnings was calculated using the formula: (Number of days of work lost × Monthly income)/30. Perceived burden of transfusion: Subjective impact on quality of life, financial strain, requirement to borrow money to afford blood products, and instances of missing scheduled HD sessions due to transfusion-related constraints. Subjective impact on quality of life and psychological burden was evaluated using unvalidated perception questions based on a 3-point Likert scale. Hemodialysis nonadherence: Number of missed dialysis sessions during the past 3 months and the documented reasons for nonadherence.

The following definition was used:

Low socioeconomic (poverty) was defined by a monthly income < $US200 associated with a poor household, or a monthly income < US$100 with an intermediate household. Intermediate to high socioeconomic status was defined by the absence of criteria for low socioeconomic status. Total cost of blood transfusion: corresponded to the cumulative financial sum paid for donor recruitment, non-medical intermediary fees, cross-matching, transportation expenses, and institutional blood bank processing fees. Recent transfusion: defined as a patient who has been transfused within the past 3 months. Anemia was defined as a hemoglobin level < 12 g/dl in women and < 13 g/dl in men.

Administrative approvals were obtained from the institutional management boards of both the Douala General Hospital and the Laquintinie Hospital. The study was conducted in strict accordance with the ethical principles outlined in the Declaration of Helsinki. All participants provided written informed consent prior to enrollment, and data confidentiality was strictly maintained using anonymized identifiers.

The data were analyzed using SPSS version 26. Continuous variables are expressed as medians with interquartile ranges (IQRs) or means with standard deviations (SDs), depending on the normality of the distribution. Categorical variables are presented as frequencies and percentages. Comparisons between groups were performed using Student’s t-test or the Mann-Whitney U test for continuous data and the chi-square test or Fisher’s exact test for categorical data. A two-sided p value < 0.05 was considered statistically significant.

Of the 225 patients screened, 168 were included. The male-to-female ratio was 1.12, with 53% (n = 89) males, and the mean age was 49 ± 12 years. The baseline socioeconomic indicators revealed widespread vulnerability (Table 1): 63.7% (n = 107) were unemployed, 65.5% (n = 110) reported a monthly income below US$200, and 69% (n = 116) had not attained a university-level education n. Furthermore, 65.5% (n = 110) of the patients were married. Poor households accounted for 42.3% (n = 71) of the patients, and low socioeconomic status was identified in 58.3% (n = 98).

Table 1. Sociodemographic and economic characteristics.

The predominant blood phenotypes within the study population were groups O+ (53.6%, n = 90) and A+ (21.4%, n = 36). The median hemodialysis vintage was 2 years [IQR: 1 - 4]. Most patients (98.8%, n = 166) underwent twice-weekly sessions, while only two patients (1.2%) received three sessions per week. Arteriovenous fistulas were the primary vascular access in 85.2% (n = 143) of cases. The mean hemoglobin level was 8.6 ± 0.77 g/dl, and 66% of patients had a hemoglobin level below 10 g/dl (Table 2).

Table 2.Hemodialysis data and anemia management.

Blood transfusion was the primary therapeutic modality for anemia management; 62.5% (n = 105) of the participants received at least one unit of blood during the preceding 6 months, culminating in an aggregate consumption of 438 blood units. More than a quarter of the patients (27.4%, n = 46) required a high transfusion burden of two or more blood units per month (Table 2).

Concerning ESAs, the use of recombinant EPO was reported by 65% (n = 109) of patients. However, among those prescribed EPO, only 62% (n = 68) reported therapeutic compliance, and 35.8% (n = 39) experienced persistent transfusion dependency. Intravenous (IV) iron therapy was administered to 59.5% (n = 100) of patients, with a self-reported compliance rate of 71% (n = 71); despite this, 42% (n = 42) required concurrent blood transfusion. Combined therapeutic regimens utilizing both EPO and IV iron were documented in 55.3% (n = 93) of patients, of whom 40.8% (n = 38) remained transfusion-dependent (Table 2).

During the 3 months preceding the study, 48.2% (n = 81) of the patients were transfused, utilizing a total of 298 blood units (186 units of packed red blood cells and 112 units of whole blood) with a mean transfusion rate of 1.23 units per month per patient. To secure blood products, 58% (n = 47) of these transfused patients relied on non-medical intermediaries, incurring a median transactional fee of US$20 [IQR: 10 - 40] per transaction (Table 3).

The structural supply chain of blood relied heavily on replacement donation (92.6%, n = 75), within which financially compensated donors represented 57.3% (n = 43), family replacement donors accounted for 12% (n = 9), and a combination of both was utilized by 30.7% (n = 23). Most replacement donors (88%, n = 66) were compensated. At the DGH, blood units sourced from external institutional blood banks incurred an additional crossmatch verification fee of US$16 per unit. Aggregating all components, the median total cost per blood unit was US$81.5 [IQR: 60 - 90] (Table 3). Additionally, the process involved a median logistical time loss of 2 days [IQR: 1 - 3] per transfusion event and a median estimated loss of earnings of US$6.70 [IQR: 3.4 - 16.7].

Table 3.Cost of blood transfusion.

*Estimated cost per unit of blood in US$.

Nearly one-third of the study population (n = 51, 30.3%) reported missing at least one scheduled hemodialysis session during the 3-month reference period (Table 4), resulting in a total of 83 missed sessions (mean of 1.62 ± 0.7 missed sessions per nonadherent patient).

Nonadherence was significantly more prevalent among patients categorized with low socioeconomic status compared to those in intermediate-to-high tiers (36.7% n = 36 vs 21.4% n = 15, p = 0.041).

Patients who underwent blood transfusion within the past 3 months also demonstrated a significantly higher rate of nonadherence compared to non-transfused peers (50.6% n = 41 vs 11.5% n = 10, p < 0.001).

Acute blood unavailability was identified as the principal reason for missing sessions, accounting for 43.1% (n = 22) of nonadherence. The frequency of this barrier did not differ significantly between low and intermediate to high socioeconomic brackets (47.2% in low, n = 17 vs 50% in intermediate to high, n = 5, p = 0.62). Vascular access complications (29.4%, n = 15), personal reasons (19.6%, n = 10), and transport constraints due to the 2025 post-electoral crisis (7.8%, n = 4) constituted the remaining documented causes (Table 4).

The subjective and economic pressure of blood acquisition was widespread among the patients. To afford blood products, 43% (n = 55 of the 128 respondents) of the patients required personal loans. The necessity to borrow money was associated with recent transfusion history (54.3% in transfused, n = 38 vs 29.3% in non-transfused, n = 17; p = 0.004) but did not vary significantly across socioeconomic strata (44.3% in low, n = 31 vs 41.4% in intermediate to high, n = 24; p = 0.81). Among actively employed individuals (n = 53), 34% (n = 18) reported missing work explicitly due to transfusion-related logistics.

Blood availability was perceived as a definitive structural barrier to medical care by 44% (n = 74) of all respondents. This perception was tightly linked to recent transfusion exposure (66.7% in transfused, n = 54 vs 23% in non-transfused, n = 20; p < 0.001) but was uniform across socioeconomic groups (48% in low, n = 47 vs 38.6% in intermediate to high, n = 27; p = 0.22). Blood unavailability was also perceived as negatively impacting quality of life (51.8%, n = 87), especially among recently transfused patients (80.3% n = 61 vs 38.3% n = 26, p < 0.001), patients with poor household wealth (76% in poor, n = 54 vs 34% n = 33, in intermediate to wealthy, p < 0.001), and low socioeconomic group (63.2% n = 62 in low vs 35.7% n = 25 in intermediate to high, p < 0.004).

The majority of patients (62.4%, n = 103 of 165 respondents) reported that the direct cost of blood negatively impacted their quality of life, a perception significantly heightened among recently transfused patients (86.5% in transfused, n = 70 vs 38% in non-transfused, n = 33; p < 0.001), but comparable among socioeconomic strata (63.3% in low, n = 62 vs 58.6% in intermediate-to-high, n = 41; p = 0.53). Similarly, the administrative and logistical procedure to locate blood was perceived as an active burden by 48.2% (n = 77 of 160 respondents) of the patients, also correlating significantly with recent transfusion history (60.5% in transfused, n = 49 vs 36% in non-transfused, n = 28; p = 0.002), and comparable among socioeconomic groups (49.5% in low, n = 46 vs 46.3% in intermediate-to-high, n = 31; p = 0.68).

Overall, 54.2% (n = 91) of the participants categorized transfusion costs as entirely out-of-pocket, and 37% (n = 62) described the logistical pursuit of blood as an acute stressor characterized by fear of missing dialysis, fear of death, donor recruitment struggles, and prolonged waiting times. Conversely, 52.4% (n = 88) noted that financial liquidity and the utilization of intermediaries were the primary facilitators of blood availability (Table 4).

Table 4.Burden of blood transfusion.

Administrative approvals were obtained from the institutional management boards of both the Douala General Hospital and the Laquintinie Hospital of Douala. The study was conducted in strict accordance with the ethical principles outlined in the Declaration of Helsinki. All participants provided written informed consent prior to enrollment, and data confidentiality was rigorously maintained using anonymized identifiers.

The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.