Sage Journals: Discover world-class research

Abstract

Background

Data quality encompasses completeness, accuracy, integrity, timeliness, and confidentiality. In many low- and middle-income countries, including Somalia, data from RHIS are often poor, limiting their usefulness for public health actions. This study assessed the quality of RHIS data and associated factors among public health facilities in the Banadir region, Somalia.

Methods

A facility-based cross-sectional study was conducted from October to December 2024 across 36 public health facilities using a multistage sampling approach. Data were collected through document reviews, interviews, and observations using PRISM-based standardized tools. Data were analyzed in SPSS version 27 after checking logistic regression assumptions. Data quality was assessed by the dimensions of accuracy (≥80%), completeness (≥85%), and timeliness (≥85%). Bivariable and multivariable logistic regression analyses identified associated factors.

Results

A total of 398 healthcare workers (59.5% female) participated, yielding a 98% response rate. Overall, good-quality data were observed in 65.3% of departments. Departments in health centers were 2.7 times more likely to report good-quality data than hospitals. Feedback, refresher training, and user-friendly reporting formats were significantly associated with better data quality.

Conclusion

Data quality across the three dimensions was scored at (65.3%). Strengthening supervision, feedback, and context-specific training can improve data reporting and management.

Keywords

routine health information systems data quality healthcare facilities Somalia

Background

Data quality is a multidimensional concept that encompasses completeness, accuracy, integrity, precision, timeliness, and confidentiality.¹ Reliable and timely data are crucial for monitoring population health, evaluating system performance, and guiding evidence-based policy and program decisions that support the achievement of health-related Sustainable Development Goals (SDGs).² A well-functioning health information system aims to routinely generate quality data that are accurate, consistent, and timely for effective health system planning and management.³

The Routine Health Information System (RHIS) is one of the six essential building blocks of health systems in low- and middle-income countries (LMICs), and serves as the primary mechanism for generating and using health data to strengthen other system components.⁴ It provides a framework for decision-making across multiple areas, including policy development, governance, regulation, human resource planning, education and training, service delivery, and system support.^1,5 In low- and middle-income countries (LMICs), routine health data represent the primary source of information for shaping health policies, monitoring programs, and improving service delivery.⁶

However, in many LMICs, RHIS remain fragmented and poorly coordinated, with widespread concerns regarding the quality, accuracy, timeliness, and completeness of their data.⁷ Weak data management, limited supervision, and inadequate communication practices have been widely reported in developing countries, contributing to poor RHIS data completeness and timeliness.⁸ The magnitude of this problem is significant, as studies from across Africa indicate data quality performance ranging from 34% and 72%, revealing substantial deficiencies.⁹ This poor quality directly leads to the underutilization of data in public health facilities, undermining effective planning and management.¹⁰

Similar challenges have been reported across African countries, where studies assessing routine health information systems have documented substantial variations in data accuracy, completeness, and timeliness. The accuracy of routine data was found to be 74%,¹ 76%,¹¹ 79%¹² of studies conducted in West Gojjam Zone, northwest Ethiopia, Hadiya Zone of Ethiopia and Nigeria, respectively. Content completeness rates were observed in studies conducted in West Gojjam Zone (70%), Hadiya Zone (83.2%), Addis Ababa (96%), Harare Zone (93%), and Rwanda (98%).^1,4,13,14 In the neighboring country of Ethiopia, the data quality is problematic for many indicators, with data quality falling below the national anticipation of 80%.⁴ Data completeness, accuracy, and timeliness varied across different regions of the country, ranging from 33% to 78%.^4,15

However, studies on HIS data quality are limited in number and scope.¹⁶ Most studies have used small-scale facility assessments and did not investigate technical or behavioral factors influencing data quality.¹⁶ Where such factors have been examined, research indicates data quality is affected by a range of technical, behavioral, and organizational factors, including staff knowledge and expertise,⁴ negligence in keeping data quality and data manipulation,¹⁷ staff motivation,^17,18 user-friendliness of reporting tools, standardized indicators,¹⁹ DQA training,²⁰ feedback,¹⁹ supervision,²¹ staff responsibility,²² and HIS data use.²³

In Somalia, the Health Management Information System (HMIS) was established between 2011 and 2012.²⁴ During this period, multiple data collection systems were implemented to gather information from healthcare facilities, many of which were managed by non-governmental organizations.²⁴ In 2017, Somalia rolled out the District Health Information Software 2 (DHIS2) across all districts, marking a major step toward improving its HIS system, and between August 2020 and December 2021, the HIS tools underwent a significant revision.²⁵ Workshops to review proposed indicators and data collection tools were conducted with the Ministry of Health (MOH) and stakeholders in Mogadishu, Hargeisa, and Garowe, aiming to refine draft tools and offer feedback to Oslo university experts.²⁶ District Health Information Software, version2 (DHIS-2), an adaptable and open-source platform with a user-friendly interface, was developed by the University of Oslo and is widely adopted for managing routine health information in more than 70 low- and middle-income nations (LMICs).²⁷

Despite these developments, limited empirical evidence exists regarding the quality of data produced by routine health information systems in Somalia.²⁸ Most available reports focus on system implementation or digital rollout rather than assessing data accuracy, completeness, or timeliness. To justify the evidence gap, we searched peer-reviewed publications and, in addition, grey literature and preprints (e.g., institutional reports, theses/dissertations, conference abstracts, and preprint servers). This approach ensured a comprehensive review of available literature, as some relevant studies may not have been formally published yet.

This study, therefore, aimed to assess the data quality of routine health information systems and associated factors among departments in public healthcare facilities, Banadir Region, Somalia. This study contributes to the field of health informatics by highlighting how routine data quality challenges in conflict-affected, low-resource settings such as Somalia, can undermine decision-support systems. It provides practical, system-level recommendations to enhance digital health reporting and strengthen the performance of health information systems.

Methods and materials

Study design and study area

A facility-based, quantitative cross-sectional study was conducted from October to December 2024 in public healthcare facilities, Banadir Region, Somalia. The initial participants consisted of 403 healthcare workers drawn from 36 public health facilities located in eight randomly selected districts: Wadajir, Kahda, Daynile, Dharkenley, Abdiaziz, Hamarweyne, Boondheere, and Wardhigley.

The study was conducted in the Banadir Region, one of Somalia’s 18 administrative regions, which serves as the country’s political, economic, and cultural center. It includes Mogadishu, the national capital city. It is geographically bordered by the Indian Ocean to the east and south, Lower Shabelle Region to the west and southwest, and Middle Shabelle Region to the north.^25,28 It is composed of 20 administrative districts which are located in an entirely urban area. The study focused on eight of these districts, chosen to represent a diverse range of healthcare environments. Banadir is home to 2,846,043 people, accounting for 13.4% of Somalia’s total population, with a gender distribution of 49.3% male and 50.7% female.²⁹

This region has the highest concentration of healthcare facilities in Somalia, with a total of 205 health facilities.³⁰ These include a range of national, regional, and district hospitals, specialty hospitals, health centers, and private clinics. In addition, about 40% of these facilities are privately owned, reflecting the region’s significant private healthcare sector.^30,31 Banadir Region also boasts the highest density of healthcare workers in Somalia, with approximately 17 core health workers (including physicians, nurses, and midwives) per 10,000 population.³⁰

Study population

The study population comprised departments within the selected public health facilities whose heads had at least six months of experience in their respective roles. Departments that were on leave, absent during data collection, or declined to participate were excluded from the study.

Sample size determination

The sample size for this study was determined using the formula for a single population proportion. Since no prior study on RHIS data quality was available in Somalia, a 50% proportion (p = 0.5) was assumed as a conservative estimate to yield the maximum possible sample size. The use of a 50% proportion (p = 0.50) for sample size calculation is a widely recommended approach when reliable prevalence estimates are unavailable, as it maximizes the statistical efficiency for a given margin of error. This method has been commonly applied in health system studies with no prior prevalence data.^32,33 A 95% confidence level and a 5% margin of error were applied. Additionally, a 5% non-response rate was included to ensure sufficient representation despite potential participant dropout. Based on these assumptions, the sample size calculation followed the formula:

n = (Z^2 * p (1 - p)) / d^2 n = ((〖 1.96) 〗^2 * 0.5 (1 - 0.5)) / 〖 (0.05 〗^() 2) n = (3.8416 * 0.25) / 0.0025 n = 384.16 .

n_final= 384 + non-response rate (5%) is equal to 403. Therefore, the final sample size used in this study was 403 participants. The sample size of 36 health facilities (33 health centers and 3 hospitals) was selected as a census of all public health facilities in the chosen districts to minimize selection bias. Given the logistical constraints and the availability of all eligible public health facilities, this number was deemed sufficient for a robust analysis of data quality across departments. The sample size was distributed across health facilities using unequal proportional allocation. Although a multi-stage cluster sampling technique was employed to ensure geographic representation, a design effect was not applied, consistent with previous RHIS data quality assessments in low-resource settings such as Tanzania.¹⁰ In these field-based studies, sample sizes were primarily guided by feasibility and available resources, rather than theoretical adjustments for clustering.

Sampling procedure

A multistage cluster sampling method was used across Banadir districts. Health facilities were defined as public hospitals and public health centers, the only public facility types in the selected districts. After selecting districts, we included all eligible public facilities (a census) to minimize selection bias, yielding 36 facilities (33 health centers, 3 hospitals). Participant sample size was allocated unequally by the number of health personnel per facility (Figure 1).

Figure 1.

Schematic illustration of the sampling procedure, 2024 (n = 403).

Variables and measurements

Dependent variable

The dependent variable was data quality (Good/Poor). Data quality was computed using three dimensions accuracy, completeness, and timeliness. For each facility, these dimensions were measured, summed, and averaged. A facility was classified as having good data quality (Yes) if it met all three minimum thresholds: ≥80% accuracy, ≥85% completeness, and ≥85% timeliness. Facilities not meeting these criteria were classified as having poor data quality (No).

Independent variables

The independent variables were categorized into three groups:

Technical factors: Presence of standard indicators, report formats, and trained personnel able to fill formats.

Organizational factors: Training, feedback, supervision, computer availability, internet access, rewards, engagement in HIS activities, performance review meetings, and data use.

Behavioral factors: Motivation, attitude, data manipulation for competition, negligence, and data quality checking skills.

Because respondents represented different departments within each facility, the technical, organizational, and behavioral characteristics were assessed at the department level.³⁴ Since the data were collected from various departments, some variability in responses was expected due to differences in workflows, roles, and staff familiarity with RHIS practices. This variability was considered during data analysis, and departmental-level analysis was employed to capture these differences. However, facility-level data quality was assessed using routine reports and records, which helped minimize any bias introduced by department-level variations.

Operational definitions

To assess HMIS data quality, three key dimensions like accuracy, completeness, and timeliness were measured, summed, and averaged to generate data quality.^13,35

Data accuracy is the extent to which reported data correspond to the original records in source documents or registers. It was evaluated by comparing figures recorded in source documents or registers with those reported to higher levels. A 10% margin of error was used to interpret the results: a ratio below 0.90 indicated over-reporting, between 0.90 and 1.10 was considered acceptable, and above 1.10 signaled under-reporting.³⁶

Completeness is the degree to which required data fields in reporting forms or registers were filled in. Completeness was evaluated by examining the number of cells in the register were left blank without a “zero” entry. If 90% or more of the cells were properly filled out, the report was considered complete.^36,37

Good Data Quality is defined as meeting the minimum acceptable thresholds across all three dimensions. Data was considered to be of good quality if it met all three criteria: at least ≥ 80% accuracy, ≥ 85% completeness, and ≥ 85% timeliness.⁶

Report Timeliness is the extent to which data reports were submitted on time, as defined by the reporting deadline. Determined by the number of reports submitted by the facility head before the deadline, the end of each month³⁸ divided by the total number of expected reports.

A data item: A single health indicator (for example, the number of ANC 4 visits, malaria cases, or vaccines) is a chosen routine health indicator used to measure the accuracy of reported data.³⁴

Data quality measurement

The three dimensions of data quality assessed in this study accuracy, completeness, and timeliness were selected based on internationally recognized frameworks. Both the World Health Organization WHO-DQR Toolkit^39,40 and the Performance of Routine Information System Management (PRISM) framework recommend these as core indicators for evaluating the performance of RHIS, particularly in low- and middle-income countries.⁴¹

To assess accuracy, a total of 36 health facilities were included. According to the Somali National HIS Strategic Plan, reviewing six to seven data elements from each health facility provides a reliable measure of data accuracy.³⁷ Accordingly, six indicators (Pentavalent 3, BCG, ANC4 visits, deliveries, pneumonia, and malaria) were selected because they are included in Somalia’s national HIS core indicator list for DHIS2 reporting^28,35 and are also among the core facility indicators recommended by the WHO DQR Toolkit.⁴² Completeness and timeliness were evaluated using a 90% threshold, based on filled report fields and timely submissions, respectively.

Data collection instrument

A structured self-administered questionnaire was adapted from the validated PRISM assessment tools (version 3.1; see Additional file 1),^16,41 with minor contextual modifications to collect quantitative data. It included sections on sociodemographic characteristics of department heads, and the organizational, behavioral, and technical factors associated with the quality of routine health information system data. In addition, observations, interviews, and document reviews guided by a standardized observation checklist were conducted to assess data quality at the departmental level. Data were collected from the respective department heads or designated representatives to ensure accurate and comprehensive responses.

Data collection procedure and data quality control

The instruments were pretested in a public health facility outside the study area using 10% of the total sample (n = 40). Based on the pretest results, adjustments were made to improve item clarity, and question sequencing. Tools were initially developed in English, translated into Somali to ensure conceptual equivalence, and then back-translated into English to verify accuracy.

A total of eight health professionals with prior data collection experience and training in data management, along with two HIS monitoring team members, were recruited from districts outside the study area to minimize potential bias. The data collectors and supervisors underwent two days of intensive training covering study objectives, data collection procedures, ethical principles, and confidentiality. The training also emphasized neutrality, accuracy, and strict adherence to data verification procedures.

Data were collected in person through on-site visits to the selected health facilities. At each visit, the study objectives were explained, confidentiality was assured, and written informed consent was obtained from facility heads and respondents. Self-administered questionnaires were completed privately to minimize response bias, while direct observation and document review were conducted to verify responses.

Supervisors conducted daily reviews of completed questionnaires and checklists to verify completeness, consistency, and accuracy. Immediate feedback was provided to data collectors to correct any omissions or errors. In addition to these reviews, regular checks were performed to ensure that data entries matched the original records, with any discrepancies addressed immediately. Incomplete tools were returned the same day for correction during the facility visit. Prior to analysis, all data were checked for completeness, cleaned, and examined for statistical assumptions relevant to the analytical models.

Data processing and analysis

The collected data were cleaned in excel, and then exported to SPSS version 27 for coding and statistical analysis. Descriptive statistics, including frequencies, percentages, means, and standard deviations, were used to summarize key variables, participant characteristics, and departmental data across public health facilities. Results were presented using tables and figures.

The dependent variable (overall data quality), was computed using three dimensions of accuracy, completeness, and timeliness as outlined in the WHO Data Quality Review Toolkit. Each dimension was scored as a percentage, and health facilities achieving an overall (accuracy, completeness, and timelines) score of 90% or above across all three dimensions were classified as having good data quality, whereas those scoring below this threshold were categorized as having poor data quality.

Because the dependent variable was dichotomous (good vs. poor data quality), binary logistic regression was employed for inferential analysis. Initially, bivariate logistic regression was conducted to identify variables associated with data quality. Variables with a p-value less than 0.25 in the bivariate analysis were considered for inclusion in the multivariable logistic regression model. The multivariable model was then used to control for potential confounding and to determine the independent predictors of routine health data quality. Results were reported using adjusted odds ratios (AORs) with 95% confidence intervals (CIs), and statistical significance was set at p < 0.05.

Prior to the final analysis, model assumptions were verified. Multicollinearity was assessed using the Variance Inflation Factor (VIF), applying a cutoff value of 10, and no multicollinearity was detected among the independent variables included in the model. The Hosmer–Lemeshow goodness-of-fit test was used to evaluate model fitness.

Results

Sociodemographic characteristics of the departments in public health facilities in Banadir region

The average age of the respondents was 30.27 years (SD ± 8.25). The majority of respondents were female (59.5%). Approximately 34.7% had 6–10 years of work experience, about half (50.5%) were nursing professionals, and 78.9% held a bachelor’s degree. Out of a total of 403 departments across 36 public health facilities in the Banadir Region, 398 departments participated in the study, resulting in a response rate of 98%. Of these, 228 (57%) departments were from hospitals, while 170 (43%) were from health centers. In terms of departmental distribution, 121 (30.4%) were from RMNCH departments, 89 (22.3%) from OPD, and 51 (13.0%) from Immunization Departments (Table 1).

Table 1.

Socio-demographic and characteristics of departments in the study on the quality of routine health information system (RHIS) data in public health facilities of Banadir region, Somalia, 2024 (n = 398).

Variables	Frequency	Percentage (%)
Name of the department
Reproductive, maternal, newborn, and child health (RMNCH)	121	30.4
Under five OPD department	89	22.3
Over five OPD department	92	23.1
Nutrition section	14	3.7
Immunization room	51	12.8
Laboratory	12	3.0
Pharmacy	15	3.7
In-patient/wards	4	1.0
Sex
Male	161	40.5
Female	237	59.5
Age in years
<=30	253	63.6
>30	145	36.4
Level of education
Diploma	41	10.3
Bachelor degree	314	78.9
Master degree	43	10.8
Field of study
Lab Technician	31	7.8
Medical Doctor	25	6.3
Midwife	32	8.0
Nursing	201	50.5
Nutrition	12	3.0
Other Field	8	2.0
Pharmacy	31	7.8
Public Health	58	14.6
Years of experience
1-5 Years	234	58.8
6-10 years	138	34.7
11+ years	26	6.5
Districts
Abdiaziz	15	3.8
Bondhere	10	2.5
Deynile	63	15.8
Dharkenley	25	6.3
Hamarweyne	81	20.4
Kahda	30	7.5
Wadajir	162	40.7
Wardhigley	12	3.0
Type of Health Facility
Hospital	228	57
Health Center	170	43

Organizational factors

In this study, 111 participants (27.9%) reported to have received feedback, while 151 (37.9%) of them had received refresher training on Health Information System (HIS) data quality within the past six months. Additionally, 185 participants (46.5%) reported receiving rewards for good work. About 125 (1.4%) of the departments had no access to an internet. Moreover, 35.2% reported receiving supervision, and 52.8% indicated that performance review meetings were conducted in their departments (Table 2).

Table 2.

Organizational factors influencing the quality of routine health information system (RHIS) data among departments in public health facilities of Banadir region, Somalia, 2024 (n = 398).

Variables	Frequency	Percentage (%)
Refreshment training in the last 6 months
Yes	151	37.9
No	247	62.1
Computer access
Yes	134	33.7
No	264	66.3
Internet access
Yes	273	68.6
No	125	31.4
Feedback
Not received	287	72.1
Received	111	27.9
Supervision
Not received	258	64.8
Received	140	35.2
Reward for good works
Absent	213	53.5
Present	185	46.5
Engagement in HIS activities
Not engaged	296	74.4
Engaged	102	25.6
Performance review meeting
Yes	188	52.8
No	210	47.2

Technical and behavioral level factors

The study found that most departments, 250 (62.8%), have a standard set of indicators. About 172 (43.2%) reported that their reporting formats are user-friendly, and 275 (69.1%) indicated that they have trained staff able to fill the reporting formats. In addition, the majority of department heads, 161 (40.5%), expressed a positive attitude toward HIS activities. However, 67 (16.8%) of respondents reported presence of data manipulation within their departments, while 215 (54%) noted the presence of negligence in HIS-related practices (Table 3).

Table 3.

Technical and behavioral factors influencing the quality of routine health information system (RHIS) data among departments in public health facilities of Banadir region, Somalia, 2024 (n = 398).

Variables	Frequency	Percentage (%)
There is a standardized set of indicators
Yes	250	62.8
No	148	37.2
Friendly format for reporting tools
Absent	226	56.8
Present	172	43.2
Trained staff able to fill out the forms
Absent	123	30.9
Present	275	69.1
Data Manipulated
Yes	67	16.8
No	331	83.2
Negligent in keeping data quality
Absent	183	46
Present	215	54
Attitude toward HMIS activities
Positive	161	40.5
Negative	237	59.5
Data quality checking skills
Good	210	52.8
Poor	188	47.2

Data quality level details

Data quality of accuracy

Out of the 398 departments for which data correctness was verified, 239 (60%) had accurate data, with hospitals exhibiting the lowest accuracy at 57% (Figure 2).

Figure 2.

Level of data accuracy among departments of public health facility types of Banadir region, Somalia, 2024 (n = 398).

Data quality of completeness

Out of 2,388 data elements reviewed for report content completeness, 1,630 (68%) were fully completed in the reporting forms. For registration content, only 49.4% of the 1,180 cases reviewed contained all the necessary information. Overall, the study found that 758 departments (32%) had incomplete data (Figure 3).

Figure 3.

Level of completeness of data among departments of public health facilities in Banadir region, Somalia, 2024 (n = 398).

Data quality of timeliness

In this study, reporting timeliness refers to the percentage of reports submitted by facility departments on or before the official reporting deadline (end of each month) out of the total number of expected reports. Regarding reporting timeliness, the study found that 161 (95%) of health center departments, and 222 (97%) of hospital departments were assessed (Figure 4).

Figure 4.

Level of data timeliness among departments in public health facilities of Banadir region, Somalia, 2024 (n = 398).

Overall data quality

Out of the participants assessed, 65.3% had good data quality. This included 52% of hospital departments and 48% of health center departments (Figure 5). Timeliness was highest at 96%, followed by accuracy at 60% and completeness at 55%.

Figure 5.

The level of overall data quality among departments of public health facilities in Banadir region, Somalia, 2024.

Multivariable Logistic Regression Analysis of Factors Affecting the Quality of Routine Health Information System Data in Public Health Facilities, Banadir Region, Somalia, 2024.

Bivariate analysis identified several factors associated with data quality, including age, facility type, presence of refresher training, internet access, user-friendly reporting tools, feedback, and supervision. In the multivariate model, facility type, refresher training, internet access, and feedback were statistically significant predictors. The findings revealed that departments in health centers were 2.7 times more likely to have good data quality than those in hospitals (AOR = 2.73; 95% CI: 1.60–4.66, p<0.001). Staff who had received refresher training were 1.8 times more likely to possess the necessary skills for ensuring data quality (AOR = 1.86, p=0.020). Conversely, the absence of internet access reduced the odds of good data quality (AOR = 0.48; 95% CI: 0.27–0.87, p=0.015). Finally, staff who received regular feedback were over eight times more likely to produce good-quality data (AOR = 8.34; 95% CI: 4.81–14.48, p<0.001) (Table4). Participants received feedback on the accuracy and completeness of the data they entered, aimed at improving data quality by identifying and correcting errors in data entry.

Table 4.

Multivariable logistic regression analysis of factors affecting the quality of routine health information system data in public health facilities, Banadir region, Somalia, 2024 (n = 398).

Variables	Data quality of the RHIS		COR (95% CI)	AOR (95% CI)	p-value
Variables	Good (N %)	Poor (N %)	COR (95% CI)	AOR (95% CI)	p-value
Age in years
<30	94(68.1)	159(62.8)	1^Reference	1^Reference
>30	44(30.3)	101(69.7)	1.357 (0.877 – 2.099)	1.422(0.884–2.288)	.170
Type of the health facility
Hospital	94 (41.2)	134 (58.8)	1^Reference	1^Reference
Health Center	44(25.9)	126 (74.1)	2.009(1.303 – 3.097)	2.726(1.596–4.656)	<0.001^*
Refresher training
Yes	78(31.6)	169(68.4)	1^Reference	1^Reference
No	60(39.7)	91(60.3)	0.700 (.459 – 1.068)	0.376(0.206–0.688)	0.001^*
Friendly format for reporting tools
Absent	92(53.5)	80(46.5)	1^Reference	1^Reference
Present	46(20.4)	180(79.6)	4.500(2.895 – 6.995)	4.567(2.729–7.642)	<0.001^*
Internet access
Yes	73(31.1)	162(68.9)	1^Reference	1^Reference
No	65(39.9)	98(60.1)	0.679(0.447 –1.032)	0.480(0.266–0.867)	0.015^*
Feedback
Not received	77(69.4)	34(30.1)	1^Reference	1^Reference
Received	61(21.3)	226(78.7)	8.391(5.126–13.734)	8.340(4.805–14.476)	0.001^*
Supervision
Not received	57(40.7)	83(59.3)	1^Reference	1^Reference
Received	81(31.4)	177(68.6)	1.501(0.979 – 2.301)	1.164(0.723–1.873)	0.063
Staff engagement
Not engaged	45(44.1)	57(55.9)	1^Reference	1^Reference
Engaged	93(31.4)	203(68.6)	1.723(1.086 –2.734)	1.286(0.495–3.338)	0.605
Data manipulated
Yes	35(52.2)	32(47.8)	1^Reference	1^Reference
No	103(31.1)	228(68.9)	2.421(1.421–4.125)	0.858(0.413–1.782)	0.680
Attitude toward HIS activities
Poor	86(53.4)	75(46.6)	1^Reference	1^Reference
Good	52(21.9)	185(78.1)	4.079(2.637–6.312)	0.498(.117–2.120)	0.346

*P value < 0.05 from multivariable analysis.

COR- Crude odds ratio, AOR- Adjusted odds ratio.

Discussions

This study aimed to assess the quality of routine health information system (RHIS) data and the factors influencing it among public health facilities in Banadir region, Somalia, using the PRISM framework. The findings provide valuable insights into the technical, behavioral, and organizational factors affecting data quality in the region. Data quality in terms of accuracy and completeness were 60% and 55%, respectively. The data accuracy in this study was 60%, which is consistent with a study conducted in Ethiopia, where the accuracy was reported as 58.1%.⁴ However, it is below the accuracy levels reported in other Ethiopian regions, such as 74%¹ in West Gojjam Zone, 76%⁴³ in Southern Ethiopia, and 89.4%⁴⁴ in Gofa Zone, Southern Ethiopia.

Somalia faces a unique challenge that contribute to the lower data accuracy observed in this study. Decades of conflict and political instability have resulted in a fragmented healthcare system with limited resources, directly impacting the accuracy of health data. The absence of comprehensive digital systems and standardized reporting methods exacerbates these issues. Health data collection in Somalia is largely manual, and the lack of adequate training for health workers leads to inconsistent and sometimes inaccurate data reporting. This challenge is compounded by the limited capacity for regular data audits and supervision.

While Somalia’s data accuracy is lower than more developed regions, it is still better than in certain Ethiopian settings, such as the 48% accuracy found in other areas.⁴⁵ This suggests that despite significant barriers, improvements in data accuracy are possible with targeted interventions, such as capacity-building, better supervision mechanisms, and gradual implementation of digital health solutions. In high-income countries like Denmark and the United States, RHIS accuracy and completeness consistently exceed 85% due to strong governance, digital integration, and accountability structures.⁴⁶ In contrast, Somalia’s fragmented administrative systems and limited digitization impede accurate reporting, resembling challenges observed in India’s less digitized regions.⁴⁷

Content completeness in this study (68%) was lower than that reported in other regions, such as West Gojjam (70%), Hadiya Zone (86%), Harari Zone (93%), and Rwanda (98%).^1,4,13,14 The lower completeness, particularly in registration records, suggests that health workers often prioritize clinical care over data documentation due to heavy workloads and limited motivation. Similar challenges were observed in East Wollega, where lower registration completeness was attributed to time pressure and limited commitment to data recording. In contrast, countries like Denmark, the United States, and Canada, with electronic health record (EHR) systems, achieve completeness rates above 90% through automated data capture and standardized formats.^46,48

The observed 96% timeliness in Banadir surpasses reported figures from Southern Ethiopia (88.42%) and Rwanda (90%),^43,49 potentially due to the region’s compact geography and enhanced facility accessibility. In contrast, Uganda’s Kampala reported significantly lower timeliness rates (44% in 2020, 65% in 2021),⁵⁰ highlighting regional disparities. This study’s timeliness is consistent with the findings in Harari Region, Ethiopia (93.7%).⁴ High-income countries like Canada and the United Kingdom maintain near-perfect timeliness through integrated digital systems.⁴⁸ These comparisons underscore the importance of infrastructure, system integration, and regional factors in determining the timeliness of health data reporting.

Our study found that 65.3% of departments have good levels of RHIS data quality, but deficiencies persist, leading to inaccurate analyses, weak service delivery, and ineffective decision-making. This may be due to a shortage of personnel responsible for data processing and management. The findings align with data from other developing countries, where data quality typically ranges from 34% to 72%.⁹ However, they are lower than findings from Ethiopia (74%).¹ Our study is also consistent with studies reporting data quality of 51.35%⁴ and 69.6%.⁵¹

The departments in health centers were 2.7 times more likely to have good-quality data compared to those in hospitals. This is supported by findings from Ethiopia regions, where data quality was found to be better at health centers.⁴ This may be attributed to the focused attention from the Ministry of Health and other stakeholders, including capacity-building efforts. Our study found that training and feedback were significantly associated with good-quality data. These findings align with a study conducted in Ethiopia, where trained staff and feedback were significantly associated with data quality, with Adjusted Odds Ratios (AOR) of 2.25 (95% CI: 1.08–4.69) and 2.48 (95% CI: 1.26–4.85), respectively.^4,52 A recent scoping review also showed that combining feedback with other capacity-building activities contributes to data quality improvement.²¹

Another predictive factor identified in our study was the format used for reporting. Departments with a -friendly reporting format were 4.5 times more likely to report high-quality data compared to those without such a format (AOR = 4.57, 95% CI: 2.73–7.64). This finding aligns with a study conducted in Dire Dawa, Ethiopia,⁵² where departments with a friendly reporting format were 2.25 times more likely to achieve good data quality compared to those without (COR = 2.25, 95% CI: 1.18–4.31). Training enhances the understanding of Health Information System (HIS)-related activities and tools, increasing familiarity with essential HIS tools such as registers and reporting formats.

Although supervision was associated with data quality in the bivariate analysis, it was not significantly associated in the multivariate analysis in this study. This finding is consistent with a study conducted in the Harari region, Ethiopia,⁴ where supervision was found to be associated with data quality in the bivariate analysis. However, it contrasts with a study conducted in Gurage Zone, Ethiopia, where supervision was significantly associated with data quality in both analyses.²⁵ One reason for this discrepancy is that, supervision is often more focused on identifying mistakes rather than offering constructive support. A more focused and supportive supervisory approach can significantly improve areas such as data recording, processing, analysis, reporting, and overall data quality management.

Policy implications and impact on health outcomes

Somalia’s health system has been deeply affected by decades of conflict, leaving it fragmented and under-resourced. Inadequate infrastructure, reliance on donor-driven mechanisms, and weak connectivity constrain RHIS performance. Health workers face high workloads, staff shortages, and limited incentives for accurate reporting.⁵³ These challenges, along with limited digitalization and weak governance, contribute significantly to deficiencies in RHIS data quality. Addressing these issues through targeted policy interventions can lead to substantial improvements in both data quality and health outcomes.

Institutionalizing RHIS training within national workforce frameworks will equip health workers with the skills to accurately collect and report data, improving data quality for informed decision-making and resource allocation, which directly enhances health outcomes.²⁸ A more skilled workforce will lead to more accurate data, helping providers identify health trends and improve service delivery. Strengthening supervision and mentorship will improve accountability in data reporting, ensuring high-quality data for better monitoring, targeted interventions, and improved health outcomes.⁵⁴ Expanding the use of digital tools like DHIS2 will streamline data collection, reduce errors, and provide real-time data, allowing quicker responses to health trends.⁵⁵ Strengthening national data governance through audits and performance management will ensure data accuracy and accountability, enabling better resource allocation and improved health services and outcomes.⁵⁶

Study strengths and limitations

This study provides valuable insights into RHIS data quality in the Banadir Region of Somalia. As the first of its kind in the country, it contributes important evidence to inform future public health initiatives. However, several limitations must be considered. The study was conducted only at public health facilities, which may limit the generalizability of the findings to all types of health facilities. The reliance on self-reported data introduces recall and social desirability biases, and non-response bias may have affected the results, particularly if facilities with lower data quality were less likely to participate. The lack of design effect adjustment and the use of the same questionnaire across facilities with different service types may have influenced the precision and contextual relevance of the findings.

Additionally, the cross-sectional design limits the ability to establish causal relationships between the factors examined and RHIS data quality. Furthermore, while the study was purely quantitative, it did not include qualitative data, which could have provided deeper insights into the behavioral and contextual factors influencing data quality, particularly in Somalia’s fragile health system. Lastly, the representativeness of the facility selection could be a limitation, as the facilities included may not fully capture the diversity of healthcare settings across the Banadir region.

Despite these limitations, the study provides a critical baseline for understanding RHIS data quality in Somalia and highlights areas for improvement in the health information system.

Conclusion

Data quality across the three dimensions was scored at 65.3%. Refresher training for healthcare professionals was inadequate, highlighting the need for context-specific training in data management and digital health tools. Health facilities should implement regular, localized training, while district health offices and the Ministry of Health should enhance supervision and feedback mechanisms to support continuous improvement. Factors like the type of facility, Internet access, and reporting tool friendliness were significantly associated to data quality. The Ministry of Health should prioritize improving digital infrastructure and Internet access to ensure health workers have the necessary tools. Development partners should support capacity-building efforts at the district and zone levels to enhance data reporting and management. Future research should use mixed-methods approaches to explore the behavioral and contextual factors influencing RHIS data quality in Somalia.

Supplemental material

Supplemental material - Data quality of routine health information systems and associated factors among public health facilities in Banadir region, Somalia: A cross-sectional study

Supplemental material for Data quality of routine health information systems and associated factors among public health facilities in Banadir region, Somalia: A cross-sectional study by Nor Haji Osman, Abdisalan Mohamed Roble, Ibrahim Mohamed Abdi, Abdiweli Mohamed Abdi, Aweis Ahmed Moallim, Osman Abubakar Fiidow, Abdirahman Mohamed Jimale, Jamal Hassan Mohamud, Abdikarim Abdi Adam, Abdirahman Ahmed Mohamud in Health Informatics Journal

Footnotes

Acknowledgements

We extend our sincere gratitude to the supervisors, data collectors, respondents, and the Banadir Regional Health Departments for their valuable contributions to this study.

ORCID iDs

Nor Haji Osman

Ibrahim Mohamed Abdi

Ethical considerations

This research has been approved by the ethics committee from the Ethical Review Board of the Somali National Institute of Health (Ref: NIH/IRB/58/DEC/2024). Before participation, all participants were fully informed about the study’s objectives, procedures, and potential benefits.

Consent to participate

Written informed consent was obtained from each participant prior to data collection. To ensure confidentiality, data were collected anonymously, with no identifying information recorded or disclosed.

Author contributions

NHO, AMR, and AMA conceptualized the study and designed the research methodology. They were responsible for data collection, data analysis, interpretation, and drafting the manuscript. OAF, AAM, and AAM conducted proofreading and provided interpretation of the results. AMJ, IMA, AAA, and JHM oversaw the overall research process. All authors read and approved the final version of the manuscript for submission

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data supporting this study are available from the corresponding author upon reasonable request.*

Supplemental material

Supplemental material for this article is available online.

Appendix

References

Chekol

Ketemaw

Endale

, et al. Data quality and associated factors of routine health information system among health centers of West Gojjam Zone, northwest Ethiopia, 2021. Frontiers in Health Services 2023; 3(March): 1–12. https://doi.org/10.3389/frhs.2023.1059611

Birabwa

Banke-Thomas

Semaan

, et al. The quality of routine data for measuring facility-based maternal mortality in public and private health facilities in Kampala City, Uganda. Popul Health Metr 2024; 22(1): 22. https://doi.org/10.1186/s12963-024-00343-z

Azim

Haake

Khatri

, et al. Routine Health Information Systems: A Curriculum on Basic Concepts and Practice 3. Measure Evaluation 2017; 3: 1–102.

Shama

Roba

Abaerei

, et al. Assessment of quality of routine health information system data and associated factors among departments in public health facilities of Harari region, Ethiopia. BMC Med Inform Decis Mak 2021; 21(1): 1–12. https://doi.org/10.1186/s12911-021-01651-2

Dehnavieh

Haghdoost

Khosravi

, et al. The District Health Information System (DHIS2): A literature review and meta-synthesis of its strengths and operational challenges based on the experiences of 11 countries. Health Information Management Journal 2019; 48(2): 62–75. https://doi.org/10.1177/1833358318777713

Sanjel

Sharma

Gurung

, et al. Quality of routine health facility data for monitoring maternal, newborn and child health indicators: A desk review of DHIS2 data in Lumbini Province, Nepal. PLoS One 2024; 19(4): 1–15. https://doi.org/10.1371/journal.pone.0298101

Hoxha

Hung

Irwin

, et al. Understanding the challenges associated with the use of data from routine health information systems in low- and middle-income countries: A systematic review. Health Information Management Journal 2020; 51(3): 135–148. https://doi.org/10.1177/1833358320928729

Asemahagn

. Determinants of routine health information utilization at primary healthcare facilities in Western Amhara, Ethiopia. Cogent Med 2017; 4(1): 1387971. https://doi.org/10.1080/2331205x.2017.1387971

Belay

Lippeveld

. Inventory of PRISM framework and tools: application of PRISM tools and interventions for strengthening routine health information system performance. Measure Evaluation, 2013.

10.

Rumisha

Lyimo

Mremi

, et al. Data quality of the routine health management information system at the primary healthcare facility and district levels in Tanzania. BMC Med Inform Decis Mak 2020; 20(1): 1–22. https://doi.org/10.1186/s12911-020-01366-w

11.

Solomon

. Assessment of quality of data and associated factors in the health management information system among health centers of Hadiya zone, Southern Ethiopia. Addis Ababa Universty, 2018.

12.

Bhattacharya

Umar

Audu

, et al. Quality of routine facility data for monitoring priority maternal and newborn indicators in DHIS2: A case study from Gombe State, Nigeria. PLoS One 2019; 14(1): 1–21. https://doi.org/10.1371/journal.pone.0211265

13.

Getachew

Erkalo

Garedew

. Data quality and associated factors in the health management information system at health centers in Shashogo district, Hadiya zone, southern Ethiopia, 2021. BMC Med Inform Decis Mak 2022; 22(1): 1–9. https://doi.org/10.1186/s12911-022-01898-3

14.

Nisingizwe

Iyer

Gashayija

, et al. Toward utilization of data for program management and evaluation: Quality assessment of five years of health management information system data in Rwanda. Glob Health Action 2014; 7(1): 25829. https://doi.org/10.3402/gha.v7.25829

15.

Fraser

HSF

Mugisha

Bacher

, et al. Factors Influencing Data Quality in Electronic Health Record Systems in 50 Health Facilities in Rwanda and the Role of Clinical Alerts: Cross-Sectional Observational Study. JMIR Public Health Surveill 2024; 10: e49127. https://doi.org/10.2196/49127

16.

Solomon

Addise

Tassew

, et al. Data quality assessment and associated factors in the health management information system among health centers of Southern Ethiopia. PLoS One 2021; 16(10 October): 1–13. https://doi.org/10.1371/journal.pone.0255949

17.

Endriyas

Alano

Mekonnen

, et al. Understanding performance data: Health management information system data accuracy in Southern Nations Nationalities and People’s Region, Ethiopia. BMC Health Serv Res 2019; 19(1): 1–6. https://doi.org/10.1186/s12913-019-3991-7

18.

Taunyane

Mpati

. Determining the Level of Health Management Information System Data Use in Southern Region of Lesotho. International Journal of Public Health and Pharmacology 2023; 3(1): 71–91.

19.

Shama

Roba

Abaerei

et al. Assessment of quality of routine health information system data and associated factors among departments in public health facilities of Harari region, Ethiopia. BMC Med Inform Decis Mak 21, 287 (2021). https://doi.org/10.1186/s12911-021-01651-2.

20.

Ahanhanzo

Ouedraogo

Kpozèhouen

, et al. Factors associated with data quality in the routine health information system of Benin. Archives of Public Health 2014; 72(1): 1–8.

21.

Ejeta

Leta

Abuye

, et al. Community health management information system Performance and factors associated with at health post of Gurage zone, SNNPR, Ethiopia. BMC Med Inform Decis Mak 2016; 16(June): 41.

22.

MEASURE Evaluation . Tools for Data Demand and Use in the Health Sector. MEASURE Evaluation Manual 2011; 177.

23.

Braa

Heywood

Sahay

. Améliorer la qualité et l’utilisation des données grâce aux ateliers d’utilisation de données: L’expérience du Zanzibar. Bull World Health Organ 2012; 90(5): 379–384. https://doi.org/10.2471/BLT.11.099580

24.

WHO . Comprehensive assessment of Somalia’s health information system 2022, 2022.

25.

III) SHSSP 2022–2026 (HSSP. 22 December 2021 1. 2022;2026(December 2021):21–22.

26.

October H. STANDARD OPERATING PROCEDURES Ministry of Health. 2010;(February).

27.

Farnham

Loss

Lyatuu

, et al. A roadmap for using DHIS2 data to track progress in key health indicators in the Global South: experience from sub-saharan Africa. BMC Public Health 2023; 23(1): 1–8. https://doi.org/10.1186/s12889-023-15979-z

28.

WHO . Comprehensive assessment of Somalia’s health information system 2022, 2022, pp. 15–16.

29.

UNFPA . Population Composition and Demographic Characteristics of the Somali People Looking towards a Brighter Tomorrow, 2016, Vol. 2.

30.

MOH Somalia. Somalia Harmonised Health Facility Assessment 2022-2023. 2024;(October).

31.

Version M. Somalia. 2025;(February).

32.

Lwanga

Lemeshow

. Sample size determination in health studies: a practical manual. World Health Organisation, 1991.

33.

Hajian-Tilaki

. Sample size estimation in epidemiologic studies. Caspian J Intern Med 2011; 2: 289–298.

34.

World Health Organization . Data Quality Assurance (DQA), 2021, Vol. 2.

35.

Guide U. PRISM Tools.

36.

Wang

Liu

, et al. Overview of Data Quality: Examining the Dimensions,Antecedents, and Impacts of Data Quality. Journal of the Knowledge Economy 2024; 15(1): 1159–1178.

37.

Sheets S. Registers & Summary Sheets USERS’ (HF) MANUAL. Ministry of Health, 2022;(January).

38.

Gesicho

Were

Babic

. Data cleaning process for HIV-indicator data extracted from DHIS2 national reporting system: a case study of Kenya. BMC Med Inform Decis Mak 2020; 20(1): 1–15. https://doi.org/10.1186/s12911-020-01315-7

39.

WHO . Data quality review. WHO Press, 2017: , Vol. NLM classi, pp. pp. 1–8.

40.

Aqil

Lippeveld

Moussa

, et al. Improved Health System Performance Improved Health Outcomes. [Internet]. Available from: https://www.cpc.unc.edu/measure

41.

Government F. Federal Government of Somalia Ministry of Health and Human Services, 2022;(March 2018).

42.

12911_2021_1651_MOESM1_ESM (4).

43.

Doka

Worku

Negeri

, et al. Assessment of data quality and associated factors in the routine health information system among health workers in public health institutions of Gofa Zone, Southern Ethiopia: A mixed methods study. Digit Health 2024; 10: 20552076241297227. https://doi.org/10.1177/20552076241297227

44.

Kebede

Adeba

, et al. Evaluation of quality and use of health management information system in primary health care units of east Wollega zone, Oromia regional state, Ethiopia. BMC Med Inform Decis Mak 2020; 20(1): 1–12. https://doi.org/10.1186/s12911-020-01148-4

45.

Schmidt

SAJ

Sandegaard

, et al. The Danish National patient registry: A review of content, data quality, and research potential. Clin Epidemiol 2015; 7: 449–490. https://doi.org/10.2147/CLEP.S91125

46.

Bodavala

. Evaluation of Health Management Information System in India Need for Computerized Databases in HMIS. … Bodalava, Takami Fellow in International Health, …. 1998.

47.

Lewis

Weiskopf

Abrams

, et al. Electronic health record data quality assessment and tools: a systematic review. Journal of the American Medical Informatics Association 2023; 30(10): 1730–1740. https://doi.org/10.1093/jamia/ocad120

48.

Karengera

Anguyo

Onzima

, et al. Quality and Use of Routine Healthcare Data in Selected Districts of Eastern Province of Rwanda. Int J Public Health Res 2016; 4(2): 5–13.

49.

Zavuga

Migisha

Gonahasa

, et al. Timeliness and completeness of monthly disease surveillance data reporting, Uganda, 2020-2021. Pan African Medical Journal 2023; 46: 2020–2021. https://doi.org/10.11604/pamj.2023.46.3.40557

50.

Ayele

Biruk

Habtamu

, et al. Data quality and it’s correlation with Routine health information system structure and input at public health centers in Addis Ababa, Ethiopia. Ethiopian Journal of Health Development 2021; 35(Special Issue 1): 33–41.

51.

Teklegiorgis

Tadesse

Mirutse

, et al. Level of data quality from Health Management Information Systems in a resources limited setting and its associated factors, eastern Ethiopia. SA Journal of Information Management 2016; 18(1): 1–8. https://doi.org/10.4102/sajim.v18i1.612

52.

Lemma

Janson

Persson

LÅ

, et al. Improving quality and use of routine health information system data in low- And middle-income countries: A scoping review. PLoS One 2020; 15(10 October): 1–16. https://doi.org/10.1371/journal.pone.0239683

53.

Dirie

Ahmed

Abdullahi

, et al. Transforming Health in Post-Conflict Somalia: Priorities from a Multi-Stakeholder Roundtable on the 2025–2029 National Plan. J Healthc Leadersh 2025; 17(September): 459–468. https://doi.org/10.2147/JHL.S541966

54.

Mohamed

George

Ali

. Challenges and Opportunities in Utilizing Secondary Data for Health Research in Somalia: A Focus on Health Management Information Systems (HMIS). Health N Hav 2025; 17(08): 921–930. https://doi.org/10.4236/health.2025.178060

55.

Ahmed

. Health Systems Governance in Somalia: An Examination of Validity, Digital Accountability, and Community Health Workforce through Mixed-Methods Research. Advances in Social Sciences and Management 2025; 3(04): 138–167. https://doi.org/10.63002/assm.304.1056

56.

Morris

Nyongesa

Sokiri

(2024) Factors influencing data quality in routine health information systems in Maridi County, South Sudan. South African Journal of Information Management, 26(1): a1856. https://doi.org/10.4102/sajim.v26i1.1856

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.37 MB

Data quality of routine health information systems and associated factors among public health facilities in Banadir region,Somalia: A cross-sectional study

Abstract

Background

Methods

Results

Conclusion

Keywords

Background

Methods and materials

Study design and study area

Study population

Sample size determination

Sampling procedure

Variables and measurements

Dependent variable

Independent variables

Operational definitions

Data quality measurement

Data collection instrument

Data collection procedure and data quality control

Data processing and analysis

Results

Sociodemographic characteristics of the departments in public health facilities in Banadir region

Organizational factors

Technical and behavioral level factors

Data quality level details

Data quality of accuracy

Data quality of completeness

Data quality of timeliness

Overall data quality

Discussions

Policy implications and impact on health outcomes

Study strengths and limitations

Conclusion

Supplemental material

Supplemental material - Data quality of routine health information systems and associated factors among public health facilities in Banadir region, Somalia: A cross-sectional study

Footnotes

Acknowledgements

ORCID iDs

Ethical considerations

Consent to participate

Author contributions

Funding

Declaration of conflicting interests

Data Availability Statement

Supplemental material

Appendix

References

Supplementary Material