Factors influencing the availability and use of electronic medical records systems in public health facilities in Uganda: a cross-sectional assessment

This section presents the empirical results of the study, structured into eight distinct subsections. The first subsection provides descriptive summary statistics outlining the demographic characteristics of the surveyed population. The second subsection details the characteristics of the participating HFs, with a focus on key dimensions influencing the appropriateness, accessibility, and operational viability of EMR systems within these institutions. Subsections three through eight offer comprehensive summaries pertaining to ICT support infrastructure, the availability and utilisation of EMR systems across the HFs, levels of EMR adoption, system integration, and the status of data warehouse implementation.

Demographic characteristics of the respondents

Table 2 shows that the highest proportion of respondents were medical records officers (40.8%), followed by health facility managers (18.9%) while information communication technology (ICT) officers were the least represented accounting for only 4.2% (11/265). In this study, due to the diverse use and understanding of the different EMR systems, the primary respondent was the ICT officer assisted by any other relevant cadres during the interview sessions while in HFs without ICT officers, the medical records officer turned out to be the primary respondent similarly assisted by any other relevant cadres. Males were the dominant gender accounting for 69.1% while females represented only 30.9% of the respondents. In terms of age, the highest age category interviewed were Millennials (76.2%) under the age bracket 28–43 years old while Gen Z (12–27 years) were as low as 4.2%.

For the education level, the largest percentage of HF staff interacting with any EMR systems were educated to a minimum of diploma level, accounting for 45.7% of the interviewed respondents, while degree holders accounted for 31.3% and postgraduates at nearly 15%. Furthermore, nearly 40% had served in the HF they represented as interviewees for about 6 years and above while those that had served in their positions within the HF between 3 to 5 years accounted for 35.1%.

The distribution exhibits adequate experience in terms of age distribution, number of years served, relevance of the cadres using the EMR systems as well as educational attainment.

Health facility characteristics and status of EMR adoption

The study sought to explore the HF characteristics related to the factors that are likely to influence the suitability of the acquisition, availability and usability of the EMR systems within the HFs to facilitate successful integration of EMR systems currently being used in HFs. These characteristics broadly relate to the location of the HF, technological factors, HF workforce, planning and budgeting for ICT infrastructure, and governance. The term integration is used interchangeably used with the term systems interoperability to mean “the extent to which any EMR systems can seamlessly communicate information among them either locally within the HF or nationally”. EMR systems integration is defined as the ability of different information technology systems and networks to communicate, to exchange data accurately, effectively and consistently and to use the information that has been exchanged [62, 63].

Availability of systems to manage patient data

Table 3 illustrates the extent to which HFs capture, store, manage and put patient data to use, albeit with patient data protection policies. The findings show that 95.8% of HFs (254/265) have systems, either paper-based, electronic, or hybrid, that manage patient data, although most HFs use both paper-based and electronic systems. Health Management Information Systems (HMIS) tools such as the maternity register book, outpatient register, and inpatient register are used to capture patient data at user points that is later aggregated and entered into the national database “DHIS2”. However, at HF level, digital EMR systems “severs” to house such data were lacking. The EMR systems available in HFs mostly focus on specific programs and these are usually funded by multi-lateral and bi-lateral donors, where the Government of Uganda and even implementing partners are mere stakeholders with limited influence regarding the technical requirements and resources [25, 64]. This partly explains why HFs using only digital patient data storage systems at HF level is very low at only 1.9% (5/265) while those using both digital and paper-based account for 94.3% (250/265).

Infrastructure, network and system requirements

Availability of computers

Availability of computers is one of the factors that can affect penetration of EMR systems as well as their integration [48, 65]. Table 4 shows that the average availability of desktop computers in HFs is 21.4 (SD = 25.06) in each HF. The mean number of functional desktop computers was 15.99 (SD = 23.66). Statistics show that most of these computers are usually donated. The study further explored the possibility of different levels of care having different numbers of desktop computers. This was performed using the Analysis of Variance (ANOVA) statistical technique, very powerful at determining the statistical significance of group mean differences among any given factor variables. Results showed that the mean availability of desktop computers in HFs significantly varied among the different levels of care (P-value = 0.000 < 0.05). Also, Tukey’s test was used to separate the means that were not individually statistically different from each other, and the test revealed that the mean differences for NR and HC4, and NR and GH for the mean number of desktop computers donated were not significantly different from each other (P-value > 0.05). This was also applied to the mean difference between HC4s and NRs for the desktop computers procured by the HF (P-value > 0.05).

On the other hand, the mean number of laptops in HFs was 12.53 (SD = 15.94) at each HF. The mean number of functional laptops was 9.54 (SD = 14.56). The study findings also revealed that the mean availability of laptops in HFs significantly varied among the different levels of care (P-value = 0.0000 < 0.05). Tukey’s test revealed that the mean difference between GHs and NRs for the mean number of laptops donated was not significantly different from each other (P-value > 0.05). This also applied to the mean difference between HC4s and GHs (P-value > 0.05).

Availability of power supply

Figure 1 shows that most HFs (95.1% of the 265) assessed are connected to the national electricity grid. Only a small fraction, specifically 13 Health Centre IVs (HC4s), reported lacking an electric power supply. Notably, all RRHs and NRHs possess standby generators that are functional and capable of supporting the entire facility during power outages. At the HC4 level, 60% of the HFs have access to a standby generator; however, only approximately 40% of these facilities possess generators with the capacity to support the entire health facility during a blackout.

Percentage availability of electric power supply N = 265

These findings suggest that, while power availability is adequate across higher levels of care, it remains a critical concern at the HC4 level. Reliable electricity is a foundational requirement for the effective deployment and sustained operation of EMR systems and other digital health interventions. Therefore, targeted interventions to enhance power reliability—such as investment in robust backup systems and infrastructure upgrades—are necessary to support the seamless integration of digital health solutions at lower-tier health facilities. Addressing these gaps will be essential to ensuring consistent service delivery and fostering digital transformation within Uganda’s healthcare system.

Figure 2 shows that the total number of HFs with installed solar power systems was 84.5% (224/265) of which 23.7% (53/224) had a solar system with the capacity to supply the entire HF. The number of HFs with solar power availability increased as the level of care descended from the NR level to HC4 level.

Availability of solar power systems in the HF

Table 5 further demonstrates that NRs and RRHs largely have consistent power availability throughout the day compared to GHs and HC4s. The power availability rate in HC4s averages between 40 and 80% a day on a scale of 24 hours. This means that HC4s and GHs should always have power back-up systems to ensure uninterrupted use of EMR systems. Only 2 HFs namely Kawempe National Referral Hospital and Mulago National Referral Hospital indicated 100% power availability throughout the whole day.

Network and internet connectivity

In fiscal year 2006/2007, the Government of Uganda (GOU) initiated the implementation of the National Data Transmission Backbone Infrastructure and e-Government Infrastructure (NBI/EGI) to enhance connectivity across ministries, departments, and agencies (MDAs). This initiative aims to provide reliable and affordable internet access to all government entities, thereby improving service delivery, communication efficiency, and information sharing. The National Information Technology Authority-Uganda (NITA-U), the agency mandated to coordinate and regulate IT services within the public sector, undertook the establishment of the NBI with an overall target of covering over 3,156 kilometers across 52 districts and connecting more than 1,300 government offices.

As of the current assessment, 603 MDAs, local government sites, and public service institutions—including health facilities and universities—have been successfully connected to the NBI [66]. However, the findings presented in Table 6 illustrate limited connectivity of HFs to this infrastructure. Specifically, only 9.8% (19/194) of Health Centre IVs (HC4s) and 43.8% (21/48) of GHs are connected, resulting in an overall connectivity rate of merely 23.0% (61/265) across the sampled HFs. Notably, the highest levels of connectivity were observed among NRHs at 100% (6/6) and RRHs at 88.2% (15/17). Geographically, health facilities in the Kampala and Lango sub-regions demonstrated comparatively higher levels of connectivity, with 58.3% (7/12) of the HFs in these areas linked to the NBI.

These findings accentuate a persistent challenge in achieving equitable internet access across all levels of healthcare provision. Although the operationalization of the NBI in the 2013/14 fiscal year led to a considerable reduction in internet costs—from an average of USD 1,200 to USD 300 per Mbps per month—many public HFs continue to face significant barriers in accessing this critical infrastructure. Consequently, the potential benefits of affordable internet connectivity, including improved digital health service delivery and integration, remain largely unrealized at lower levels of care. This highlights the need for deliberate policy and investment efforts to extend NBI connectivity to underserved health facilities, thereby promoting digital inclusion and strengthening health system performance.

As can be seen in Table 7 about HFs connected to NBI; the average number of departments varied significantly depending on the level of care (p-value = 0.0035 < 0.05). On average, the mean number of departments was 14.91 (SD = 9.08) of which on avarege12.97 (SD = 13.11) departments were connected to the NBI. At the NR level, the average number of departments was nearly 30 (SD = 7.22) of which 22.00 (SD = 14.04) departments were connected to the NBI.

ICT structure, governance, policies and staffing

The recent staffing establishment approved by the Ministry of Public Service provided for the inclusion of ICT positions within health facilities (HFs) up to the HC4 level. The move was aimed at supporting the ongoing digitization of healthcare services within Uganda’s health sector. Despite this provision, the study revealed a critical gap in the availability of functional ICT units across many HFs, particularly at the GH and HC4 levels. According to the data presented in Table 8, only 2 out of the 194 HC4s reported having an ICT unit, each staffed with one Assistant ICT Officer as stipulated. Similarly, only 2 out of the 48 GHs had an ICT unit, with just one GH reporting the presence of an ICT officer.

The situation, however, improves at higher levels of care. Nearly all NRHs reported having ICT units in place, while 13 of the 17 RRHs had established ICT units, though three of these units were not staffed. It is important to note that a considerable proportion of ICT support within HFs is provided by personnel deployed by Implementing Partners (IPs), who are typically recruited to support specific health programs such as HIV and tuberculosis care.

In addition, the findings indicate that only about 30% of HFs reported having a dedicated budget for ICT-related services. However, these budgets were often limited in scope, primarily allocated for basic needs such as internet connectivity and the procurement of small-scale ICT equipment. Disaggregation by level of care revealed that 66.7% of NRs, 58.8% of RRHs, 43.8% of GHs, and 22.7% of HC4s had budgets dedicated to ICT services. These findings highlight a pressing need for more comprehensive investment in ICT infrastructure and staffing at all levels of care to enhance the sustainability and efficiency of digital health initiatives in Uganda.

Data protection and privacy

Table 9 presents findings relating to this thematic area. It was noted that on average 35% (93/265) of the hospital managers were aware of the Data Protection and Privacy ACT 2019. The highest percentage was at RRH level (82.4%, 14/17) while the lowest percentage was at HC4 level (26.8%, 52/194). The study further explored the extent to which awareness of the Data Protection Act statistically varied across the different HF levels of care. It was found that there was a statistically significant linear relationship between awareness about the data protection ACT and the level of care (p-value = 0.0000 < 0.05). This means that HF managers in highly specialized HFs are more likely to be aware of the data protection ACT compared to their counterparts in lower-level HFs. Additionally, the same test statistic did not demonstrate any significant departure from this linear trend (p-value = 0.2510 > 0.05).

On the other hand, only 2 HFs were registered with the data protection office as required by the ACT. Finally, the proportion of HFs with a data protection officer was only 9.1% (24/265); 15 out of 194 HC4s, 2 out of 48 GHs, 4 out of 17 RRHs, and 3 out of 6 NRs Using the Cochran-Armitage chi-square test, we observed a linear association between having a data protection officer and the level of care, where higher level HFs are more likely to have a data protection officer than their counterparts (p-value = 0.0026 < 0.05).

There is a need by the Ministry of ICT and NITA-Uganda to popularise the Data Protection and Privacy Act to ensure that data relating to patients and or other medical records are collected within the established regulations and in compliance with the acceptable national standards. As portrayed in this study, whereas HFs and implementing partners continue to collect such sensitive information, with the lack of such crucial officers in the HFs to regulate the standards and accessibility, chances are high that most data is collected outside of the set national laws and regulations.

In 2023, the Ministry of Health (MoH) in Uganda issued a privacy notice aimed at guiding the collection and use of data within electronic health information systems, with the primary objective of safeguarding personal data and maintaining user trust [67]. This study investigated the extent to which hospital managers, key custodians of institutional policies and procedures, are knowledgeable about and utilise the Uganda Health Data Protection, Privacy, and Confidentiality Guidelines (2023). The findings, as illustrated in Fig. 3, indicate that only a minority of hospital managers (34.7%, 92/265) were aware of the guidelines. Awareness varied by level of care, with RRHs recording the highest awareness (52%, 9/17), followed by NRHs at 50% (3/6), GHs at 43.8% (21/48), and HCIVs at 30.4% (59/194).

In terms of data protection mechanisms, the study revealed that most HFs employed physical access control measures (97.7%) and electronic access control measures (92.5%) to restrict unauthorised access to patient information. Pearson’s chi-square test was employed to assess whether the use of various data protection measures differed significantly across levels of care. The analysis showed no statistically significant association between level of care and the use of physical access controls (P-value = 0.523), electronic access controls (P-value = 0.309), or data encryption (P-value = 0.081). However, compliance with data protection regulations significantly varied by level of care (P-value = 0.005). These findings suggest that while most facilities have implemented pragmatic strategies to secure patient data, awareness and formal policy implementation remain inconsistent. To ensure robust data governance, the dissemination and institutionalization of national data protection policies must be enhanced across all levels of the health system.

Landscape of electronic medical records systems availability and integration

For the country to achieve seamless information flow and interoperability of EMR systems in the healthcare system, coordinated modalities are integral in all spheres, including capacity building of users, deliberate, strategic and focused deployment of EMR systems as well as change management [68]. Interoperability of systems assists in predicting adverse health events, proper planning and allocation of health sector resources, provide and guide patient centred healthcare services, facilitate evaluation of healthcare interventions, and ensure accuracy and accountability of health commodities. The response variable in this study was measured on two fronts: 1) the number of EMR systems in each HF, and 2) whether the system (s) was in use.

Availability of EMR systems in HFs

The current public health sector landscape is characterised by many EMR systems, often deployed in uncoordinated approaches and working in silos. The study sought to establish the current list of EMR systems being used in public health facilities. All HFs had at least one EMR system (s) in place. Since the study focused on HC4s through NRs, at bare minimum each HF was expected to have a client self-service portal (CSSP/NMS+) and use it to order for essential medicines from National Medical Stores (NMS). Also, HFs were expected to have DHIS2, a central repository system used to capture aggregated data at the HF level, store, and be accessed at the national level.

Table 10 shows that on average, each HF had between four and six EMR systems (mean = 4.81, SD = 1.41), and the difference in the average number of EMR systems between levels of care was statistically significant (p-value = 0.0108 < 0.05). Only GH & HC4; and HC4 & RRH means were statistically significantly different from each other.

Distribution of EMR systems deployed in HFs

The MoH, working with partners, has been deploying EMR systems to facilitate digitisation of healthcare services. It should be noted that legacy systems like DHIS2 have been in existence longer compared to those whose deployment started a few years ago. Table 11 illustrates the current distribution of EMR systems and the levels of care where they are predominantly deployed. CSSP/NMS+ is an ordering system used by all HC4s, GHs, RRHs, and NRs to send their orders to NMS, a national warehouse mandated by GoU to distribute medicines and health supplies to all public HFs. This system was found in all the surveyed HFs. DHIS2 is a legacy system managed by the division of health information (DHI) of MoH.

The system was deployed by MoH as a national repository database to strengthen routine reporting on EMHS use, EMHS ordering, and patients’ medical records. This system is supposed to be used by all the HFs, including private HFs. The findings revealed 100% (265/265) availability of DHIS2 in public HFs. Other dominant EMR systems currently deployed in public HFs include Uganda EMR (86.0%, 228/265), Rx solution (68.3%, 181/265), A-LIS (55.1%, 146/265), and eAFYA (16.6%, 44/265). Some of these systems target specific disease programs.

The Human Resource Information System (iHRMIS) is national human resources register that was embraced in Uganda by the capacity project in 2005, although the management of this system was later transferred to the USAID Uganda capacity program (UPC) in 2009 [69]. The IHRMIS was first adopted by the Uganda Nurses and Midwives’ Council (UNMC) in 2007 [69]. The central MoH, department of human resources was enrolled into iHRIS in 2008. During that period, HRMIS was being used to provide a platform for reporting on the health workforce as well as analytics and to enhance the ICT infrastructure [69, 70]. The Ministry of Public Service has started deploying this system in public HFs and per the findings in this study, around 4.5% of high-level HFs have been enrolled.

Uganda EMR is an EMR system used to collect data on HIV and Tuberculosis (TB) patients, although the scope was later expanded to include maternal and child health (MCH) services. Uganda EMR, a customisation of OpenEMR, was first piloted in Uganda’s HFs in 2011, covering a total of 20 pilot HFs at that time. The system is supported by METS Program at Makerere University School of Public Health [71]. Uganda EMR was deployed in 1300 public HFs by 2022, and it’s currently estimated to have a presence in about 2689 ART sites, including selected health centre IIs (HC2s) [71, 72]. The study found noticeable deployment and use of Uganda EMR in the surveyed HFs, being used to capture, store, and manage HIV and TB patients’ data at the point of care. The EMR system also has the capacity to integrate with other systems. Currently, it is integrated with the electronic-based surveillance system (eCBSS) and DHIS2 among the legacy systems. eCBSS was originally introduced to manage TB patient diagnosis and treatment, however, this was alienating it from other related conditions and thus the need to interoperate with the Uganda EMR to have information exchange on TB and HIV medical records. It was also proposed to have Uganda EMR integrated with A-LIS as well as with eAFYA especially at RRH level where the two systems co-exist [73].

EMR systems functionality

Table 12 illustrates the current state of use of EMR systems deployed in public HFs. RASS, Lab Expert, EIDSR, and Clinic Master are all in use in the HFs where they are existent. DHS2 (98.8%, CSSP/NMS+ (97.4%), Uganda EMR (96.5%), and A-LIS (95.2%) are among the EMR systems with dominant use in HFs. RX solution had the highest number of HFs (23/180) where it was deployed but not in use.

Number of EMR system users within the HF

The number of any EMR system users may vary depending on several circumstances, including technical capabilities and functions, staffing levels within the HF, targeted medical programs, training and capacity building of users, and level of care, among others. The findings in this study indicated varying mean numbers of users depending on the EMR system as well as the level of care. Table 13 shows that eAFYA has the highest mean number of users in each level of care where it is deployed and in use. The differences in the mean number of EMR system users between levels of care were statistically significant for most of the systems (p-value= < 0.05). This means that when considering the integration of EMR systems, one of the drivers is the availability of computers for use by the system users; therefore, when providing computers, the average number of EMR system users by level of care should be considered to avoid over supply or under supply of computers. Furthermore, some of these systems have modules like the prescribing module, dispensing module etc., that accommodate several points of care (POCs) with many users, therefore, some EMRs are likely to have more users than others depending on the specific system modules and acceptability.

EMR system interface

Electronic information management systems are designed in such a way that simplifies the access, use, and friendliness to users. This can easily boost acceptability among users. This study explored the extent to which HF EMR system users rank the ease of use of the EMR systems interfaces. This was ranked using a likert scale of 1 to 5, where 1 represents poor and 5 represents very good. Table 14 shows that cumulatively, Uganda EMR (93.5%), CSSP/NMS+ (92.3%), DHIS2 (91.6%), were described as EMR systems with good user interface. Other EMR systems above the 80% mark include A-LIS (86.6%), RASS (86.2%), EIDSR (84.6%), and Clinic Master (83.3%). On average, all EMR systems have a good, easy to use, and appealing user interface.

EMR system is bug/error-free, and all functionalities are working smoothly

Respondents were asked to rate the extent to which EMR systems deployed in their HFs meet their expectations in terms of smoothness in the functionality and use of specific EMR systems. The findings in Table 15 indicate that RASS (86.2%) LAB expert (85.7%), DHIS2 (85.2%), EIDSR (84.6%) and A-LIS (82.4%) rarely develop errors, and their functionalities are working smoothly.

EMR system flow and use is understandable to the user with minimal ICT skills

The findings in Table 16 indicate that CSSP/NMS+ (79.2%), A-LIS (76.1%), HRIS (72.7), and RASS (72.4%) have functionalities that are easily understood, where anyone with a minimum ICT skill set can easily navigate through.

Level of satisfaction with the EMR system deployed in the HF

The study explored the extent to which EMR systems users were satisfied with the performance of such systems. This was measured on a scale of 1 to 10, where 1 indicated least satisfied and 10 indicated most satisfied. The results in Table 17 show that users are most satisfied with DHIS2 with a mean score of about 83% (mean = 8.25, SD = 1.16) followed by RASS with a mean score of 80% (mean = 8.03,SD = 1.21), CSSP/NMS+ at 79% (mean = 7.86, SD = 1.29), Uganda EMR at 78% (mean = 7.94, SD = 1.18) and EIDSR at 78% (mean = 7.77, SD = 1.24).

The results generally imply that the level of confidence HFs have in most EMR systems is demonstrated moderate to high levels of confidence users have in these systems. If all challenges are well addressed, the results indicate smooth adoption of the integration concept by the HFs. The findings in Table 17 further revealed that the level of user’s satisfaction significantly varied depending on the system (p-value < 0.05), implying that EMR systems user’s satisfaction was more inclined to some specific systems compared to others.

Deployment of EMR systems in HFs and system support

Several EMR systems currently exist in the HFs. Most HFs have more than three systems being used. However, the deployment process remains unclear. HFs indicated lack of involvement in the process of deploying EMR systems in HFs, while at the national level, sector-wide consultations are limited. Few stakeholders are involved in the process. MoH should “Develop and ratify EMR framework that guides all stakeholders on how to select, develop and deploy EMRs in the country to ensure value for stakeholders”. (KII2)

Figure 4 shows that MoH and IPs play a pivotal role in supporting the operational status of the EMR systems when systems need any upgrading and troubleshooting. “Support requests are handled at MOH through different channels like WhatsApp, email, phone calls etc. We are currently working on a ticketing system”. (KII2)

Efforts towards the integration of EMR systems

The MoH has a technical working group responsible for the integration/interoperability and standardization of the EMR systems, national registries and the legacy systems. The findings indicate that Uganda is fostering EMR integration through training and capacity building and strengthening the development of data warehouses. As at the time of this study, the following integrations had been implemented:

• Uganda EMR was integrated with eAFYA, especially on HIV patient data.

• Uganda EMR and clinic master were integrated with Lab expert and A-LIS.

• eAFYA is integrated with lab expert.

The integration of digital systems outside the health sector is managed by NITA-Uganda and implemented through the UgHub, for example, integration with NIRA and URA. MoH is gradually phasing out legacy systems. Every HF upgraded to the current implementation of digital tools are phased out of the legacy systems.

Generally, the integration process is already happening, although being handled in phases and taking different stages depending on several factors including the modules and features of specific EMR systems, degree of HF coverage and penetration in some EMR systems, availability of servers in the HFs, and knowledge among health facility staff to use the deployed EMR systems in the HFs.

Second, the study observed varying technical capability gaps in EMR systems with varying interest in terms of scope: For example, there is no standard EMR system that currently handles patient data as well as commodity management.

Challenges affecting use of EMR systems in the HFs

The Uganda health information and digital health strategic plan 2020/21–2024/25 illustrated a multitude of challenges that were limiting the acceleration of use of digital platforms in the health sector, including human factors such as lack of skills; technological factors such as system capabilities, multiplicity of duplicate and uncoordinated health information platforms; and lack of standard operating procedures (SOPs) [5]. This study found that many of these challenges limiting the adoption and widespread use of EMR systems still exist. More particularly, the following challenges indicated in Table 18 were found prevalent.

Foundational ICT training and capacity building

ICT knowledge among the health workforce is one of the factors that can easily accelerate the adoption and use of EMR systems in the HFs. When the expected users are computer illiterate, digital transformation may turn out to be difficult to achieve because of the resistance from the users. This sometimes may require change management programs carried out among the health workers. This study explored the computer literacy levels among health workers. This was measured on a scale of 1 to 5, where 1 indicated very illiterate and 5 indicated excellent. Results in Table 19 indicate that pharmacists (mean = 4.1 SD = 0.66) have a better understanding of computer use among the health workforce compared to other cadres. This was closely followed by medical records officers (mean = 4.1, SD = 0.55). Other cadres with relatively high computer skills include store officers (mean = 3.9, SD = 0.61), consultants (mean = 3.9, SD = 0.57), medical officers (mean3.9, SD = 0.53), dispensers (mean = 3.9 Sd = 0.68), and clinical officers (mean = 3.6, SD = 0.67). Nurses, midwives, and nursing assistants are characterized by low computer literacy levels. The findings imply a smooth acceleration of EMR use in the case of integration. Most cadres that ideally should interact with EMR digital platforms right from EMHS ordering, requisitioning, issuance, prescribing, dispensing, up-to-reporting have the necessary ICT knowledge to use computers.

A linear-by-linear chi-square test was used to test for the existence of a statistically significant association among computer literacy and healthcare workforce by level of care. The outcome variable was the level of computer literacy while the level of care was treated as a factor variable. Both variables exhibited a natural ordering. Results show that the level of computer literacy between the different levels of care does not necessarily depend on the level of care where the healthcare staff is deployed (p-value > 0.05). This means that regardless of the level of care where a medical officer or a clinical officer is deployed, they will adequately use any EMR systems in place.

Data warehouse

A Data Warehouse is a central repository or platform of integrated structured or semi-structured data from one or more sources for purposes of analytics and reporting. In December 2023, the MoH was reported to have embarked on a project to develop a national data warehouse for all systems to allow easy sharing of data, visualization, strengthening of surveillance among others [74]. The following activities were being undertaken: requirements analysis, data warehouse architecture design, purchase and installation of the data warehouse equipment, data warehouse data model design, data warehouse pipelines development, data warehouse security design and implementation and jasper report & power BI training [74]. Once completed, this will accelerate and facilitate the integration of EMR systems and allow end-to-end information storage, exchange and use at national, sub-national and at HF level. Currently, the National Data Warehouse is receiving data from eCHIS, EMR, and DHIS2 (HMIS, EIDSR).

Model estimation

This study used a Poisson regression model facilitated by the class of generalized linear models to determine the factors influencing the current deployment, availability, and use of EMR systems in public HFs in Uganda. The interpretation of Poisson regression coefficients depends on the interest of the study. In this paper, the interpretation is based on the expected number of EMR systems in each HF, implying a factor change in the rates given a set of explanatory variables using Incidence Rate Ratios (IRRs) (Table 20).

Technological factors

The study findings in Table 20 revealed that the nature in which patients’ data are managed in HFs has a significant influence on the penetration of EMR systems in HFs. HFs using both paper and electronic systems were 1.18 times more likely to have multiple EMRs compared to HFs using only paper-based systems (p-value = 0.027 < 0.05). It was further revealed that HFs whose average downtime is more than one month are 1.38 times more likely to have EMR systems compared to HFs with EMR systems whose average down time is less than one hour (p-value = 0.000 < 0.05).

Organisational factors

Results in Table 20 indicate that we are 36% more likely to observe EMR systems at RRH level compared to HCIVs (p-value = 0.000 < 0.05). GHs (12%) were more likely to have multiple EMR systems compared to HCIVs (p-value = 0.041 < 0.05). This is attributed to the substantial number of departments, programs and specialty clinics at RRH level, and most of the programs have at least deployed an EMR system to manage patient data, inventory and the clinical processes specific to these programs.

Furthermore, having many medical departments within the health facility was positively associated with having many EMR systems (p-value = 0.020 < 0.05), where a one unit increase in the number of departments was associated with a 1% increase in the mean number of EMR systems in a HF. The findings revealed that HFs that provide funding towards ICT activities were 1.14 times more likely to have EMR systems in use.

Environmental factors

Among environmental factors, HFs where implementing partners provide internet support were 16% more likely to have many EMR systems in use compared to those where system users provide internet (p-value = 0.002 < 0.05).

The study included one environmental factor in the model estimation process. Other environmental factors such as availability and awareness of data protection and patient data privacy were assessed at bivariate level. These were not included in the model estimation process to preserve model stability. Furthermore, environmental factors such as procedures to guide the adoption of EMR systems and efforts towards EMR systems integration were assessed through KIIs at national level (sects. 3.3.9 & 3.3.10).