INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
Data Management Practices, Digitization, and Sustainability of  
Community Health Programmes in Nairobi County, Kenya  
Margaret Njeri Kabue*., Prof. Raphael Nyonje., Prof. Dorothy Ndunge Kyalo., Dr. Kipkorir Michael  
Chirchir  
University of Nairobi  
DOI: https://dx.doi.org/10.47772/IJRISS.2025.91100465  
Received: 04 December 2025; Accepted: 10 December 2025; Published: 18 December 2025  
ABSTRACT  
Monitoring and evaluation data from community health programs are essential for their planning, improvement,  
success, and sustainability. These data can inform implementers, policymakers, beneficiaries, and funders about  
the progress or lack of it, prompting them to take appropriate action at any stage of the project life cycle. This  
study examines the moderating influence of digitization on the relationship between data management practices  
and sustainability of community health programmes. The study is anchored on Technology Acceptance Model  
(TAM). A positivist philosophical paradigm and a cross-sectional descriptive survey research design to generate  
quantitative primary data. A Partial Least Squares – Structural Equation Model (PLS-SEM) priori sample size  
calculator was used to determine a sample size based on all observable variables, including measurement  
indicators for the moderating variables. The sample comprised 190 community health promoters in Nairobi  
County. The response rate was 83%. Data analysis was performed using SmartPLS 4. The study results affirmed  
that the impact of data management practices on the sustainability of community health programmes decreases  
with digitalisation. The results resonate with the TAM, which posits that the adoption of innovations, such as  
data management automation, is influenced by perceived benefits and ease of use for users. This study  
recommends a hybrid data management approach that combines traditional and automated practices to address  
the transition and the digital divide. The findings of this study can also be used by policymakers and programme  
implementers to design programmes with greater specificity about where to invest when introducing or scaling  
up data digitisation, as well as to address contextual factors that contribute to the sustainability of community  
health programmes.  
Keywords: Data management practices, Digitization, sustainability of community health programmes  
INTRODUCTION  
In the effort to achieve health for whole populations, governments have increasingly focused on community  
health. Perhaps establishing structures at the household level will help reach the entire population. Developing,  
expanding, and strengthening community health programmes is argued to be a promising approach to achieving  
health for all (Perry et al., 2021). Community Health Programmes (CHPs) are also associated with preventing  
disease progression, reducing healthcare costs, and promoting the population's wellbeing (Khatri et al., 2024).  
The sustainability of Community Health Programmes (CHPs) is therefore crucial to the communities that benefit  
from them and to those who implement and fund them. The implementation, scale-up, and sustainability of these  
programmes seem to depend on data that guide implementers and funders in decision-making and the selection  
of best practices (Tshuma et al., 2024; Åhlfeldt et al., 2023; Habte et al., 2022; Kirk et al., 2019).  
Studies in this area have highlighted that data quality from community health is insufficient (Lee et al., 2021);  
thus posing a need to invest in data quality to support decision-making on the sustainability of CHPs. Similarly,  
digitization affects the effectiveness of data management and can play a catalytic role in how the interventions  
are embraced, adapted, or abandoned (Chirambo et al., 2018). This study was anchored on the Technology  
Acceptance Model (TAM). TAM posits that acceptance or rejection of technology is influenced by perceptions  
of its usefulness, ease of use, and, ultimately, the attitude formed around its adoption (Nyimbili & Chalwe, 2023).  
Page 5900  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
In Kenya, CHPs are supported by a legislative framework comprising the Primary Health Care Act, the Digital  
Health Act, the Facility Improvement Financing Act, and the Social Health Insurance Act, as stipulated by the  
Community Health Units in the Universal Health Coverage Report (2020). Kenya is committed to achieving the  
global target of Universal Health Coverage (UHC) by 2030, adopting a community health strategy to help access  
services for hard-to-reach and marginalized populations (Kenya National Bureau of Statistics, 2020). This  
resulted in significant investment in establishing community health service infrastructure, supported by  
government-led legal reform through the Ministry of Health.  
An evaluation of community health services in Kenya in 2018 revealed a dire need for investment in CHPs  
(Ministry of Health, 2020). This evaluation reported three significant gaps in community health services. First,  
a 41% deficit in Community Health Units (the smallest unit of a Community Health Programme) was identified,  
reflecting community health service coverage. Second, a 17% gap in the number of recruited Community Health  
Promoters, who are critical to delivering community health services, including data management. Third, an 85%  
gap in the number of Community Health Assistants (CHAs) who are crucial in the supervision of Community  
Health Promoters. In addition to coverage, the quality of data generated by these community health programs  
has been questioned, posing a challenge to their use in decision-making at various levels (Regeru et al., 2020).  
The studies linking DMP and the sustainability of CHPs are hindered by methodological gaps that this study  
addressed. The study by Donessouné et al. (2023), which aimed to assess the sustainability of TB, Malaria, and  
HIV programmes funded by the Global Fund, used a small, convenience sample, making generalizability  
difficult. Furthermore, the study was conducted three years after the intervention, which could have introduced  
memory bias into the findings. Similarly, a study by Mehra et al. (2020) aimed to identify community- and  
institutional-level facilitators of programme sustainability, using a qualitative approach. This approach limited  
the sample size and scope, resulting in a lack of generalizability of the results to other settings. Although a study  
by Shanmuganathan et al. (2022) focused on non-communicable diseases in Malaysia and employed a mixed-  
methods study design, the data analysis relied on descriptive statistics, which may not provide the rigor required  
in research. The methodological gaps identified in the studies by Donessouné et al. (2023), Shanmuganathan et  
al. (2022), and Mehra et al. (2020) were addressed in this study. The sample for this study was obtained through  
stratified random sampling. Further, the Partial Least Squares-Structural Equation Modeling (PLS-SEM)  
technique was used to establish the sample and analyze data. This allowed simultaneous analysis of  
measurements and structural models and supports the exploratory nature of this study, as supported by Sarstedt  
et al. (2021) and Ramli et al. (2018). The choice of research methodology adopted by this study, the  
generalizability of results, and the rigor were guaranteed.  
Several studies have been conducted in urban and pre-urban settings of Low-Income and Middle-Income  
countries, including Kenya, with compelling evidence of the influence of program and contextual-level factors  
on performance of CHPs (Ogutu et al., 2021). However, the focus of these studies is limited to program  
performance and not sustainability. This study extended beyond CHP performance to their sustainability, while  
considering the unique environmental factors in these settings. The studies that came closest to discussing the  
question for this research are from other countries; the studies by Donessouné et al. (2023), Shanmuganathan et  
al. (2022), and Mehra et al. (2020) were carried out in Burkina Faso, Malaysia, and the United States,  
respectively; hence, pausing a contextual gap that this study addressed. Mremi et al. (2022) found a positive link  
between continuous feedback and sustainability. However, the study was conducted in Sweden's social care  
sector.  
Another contextual gap identified from the literature aligns with the scope, breadth, and depth of CHPs in Kenya.  
There are fourteen standard areas of focus (Ministry of Health, Division of Community Health Services,  
Republic of Kenya, 2021). The CHPs in Kenya focus on covering a broad range of health issues at the community  
level. Most studies on the performance or sustainability of CHPs concentrate on a single disease or health  
condition (Donessouné et al., 2023; Mehra et al., 2020; Regeru et al., 2020). This fails to recognise the full scope  
of data handlers’ data management work; as a result, the view is limited and may not be representative. This  
study aims to establish the influence of DMPs on the sustainability of CHPs by addressing all the program focus  
areas stipulated by the Government of Kenya.  
Page 5901  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
Based on these existing conceptual, contextual, and methodological gaps, this study potentially addressed the  
knowledge gaps. Therefore, it sought to answer the research question: What is the moderating influence of  
digitization on the relationship between data management practices and the sustainability of community health  
programmes in Nairobi County, Kenya?  
LITERATURE REVIEW  
The Technology Acceptance Model (TAM), developed by Davis, Bagozzi, and Warshaw in 1989, is one of the  
most widely used models for explaining the adoption of technology across various fields (Saade et al., 2011).  
According to Davis (1989), an individual will adopt technology or innovation based on how they perceive its  
benefits and ease of use, leading to an attitude that determines whether they will adopt it (Nyimbili & Chalwe,  
2023). Therefore, the TAM model helps understand how digitization or the use of technology is accepted and  
adopted and attempts to explain why digitization will be embraced or rejected in managing a firm, organization,  
or program.  
While digitization has the potential to make data management functions easier, faster, and less tedious, it is not  
always readily adopted, as it is perceived as new, unexplored, and less understood by a portion of its intended  
audience and users (Nyimbili & Chalwe, 2023). Other factors, such as training, user participation in design and  
implementation, and the system’s features, could also influence the acceptance or rejection of digitization, as  
noted by Silva (2015), who further explores how system design can influence the acceptance or rejection of the  
technology. The primary data handlers in CHPs lack adequate participation in the design of, and the use in a  
range of other essential data management functions (Kenya Community Health Policy 2020-2030).  
The applicability of the TAM across contexts can be contested, especially in low-resource settings and in the  
face of limited technological infrastructure (Zaidi et al., 2020; Bagozzi, 2007); the model fails to account for  
other aspects that may be institutional or environmental in nature. Despite the reported high penetration of  
mobile connectivity in Nairobi, a significant portion of the population resides in slums with very low  
socioeconomic status, compounded by insecurity; these conditions are likely to influence the adaptability of  
digitization, a factor that TAM has not addressed. This study seeks to establish the influence of digitization on  
the relationship between DMPs and the sustainability of CHPs.  
The relationship between data management practices, digitization, and sustainability of CHPs can be argued  
based on TAM (Feroz et al., 2020) or systems theory. Systems theory suggests that digitization can be applied  
to various interconnected and interdependent aspects of data management to enhance system efficiency  
(Donessouné et al., 2023). This means that different parts of data management practices, such as collection,  
storage, and analysis, can be digitized to influence the overall data quality. While such practices can be digitized  
independently, systems theory suggests that to achieve the best results, each subset of a system needs to be  
addressed. Drawing from the TAM, individuals will embrace digitization only if they find it valuable and easy  
to use. Digitization can make data management more efficient, improve quality, encourage adaptive  
implementation, enhance evidence-based and timely decision making of the community health programmes,  
thus facilitating their sustainability (Dillip et al., 2024; Kansiime et al., 2024; Kaboré et al., 2022; Owoyemi et  
al., 2022). However, this depends on its availability and the ability of data handlers to utilize it effectively. This  
section presents the link between data management practices, digitization, and the sustainability of CHPs. This  
study argues that digitization in data management has the potential to enhance data quality, which is essential  
for the sustainability of CHPs.  
The timeliness and completeness of community health data appear to be exacerbated by environmental factors  
such as long distances, poor infrastructure, and unhygienic working conditions (Kirk et al., 2021). Some studies  
attribute the use of MobileApps to improved access and speed to collect, retrieve, and share data; enhanced  
reporting and monitoring mechanisms of CHP work; reduced workload, advanced personal development, and  
improved the status of Community Health Promoters in the community they work in (Njororai et al., 2021; Zaidi  
et al., 2020). There is evidence that the use of mobile devices to collect data and communicate on essential  
services and processes among community workers has addressed the issues of long distances, poor infrastructure,  
and sometimes language or cultural barriers, as the data handler does not need to collect data in person (Yang et  
Page 5902  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
al., 2021). There was value addition in data quality with digitisation, especially in timeliness, completeness, and  
consistency of reports (Unkels et al., 2023; Kirk et al., 2021).  
On the contrary, a study by Muinga et al. (2020) and Zaidi et al. (2020) found no significant differences in the  
quality of data collected using electronic gadgets, paper, pen, and blended approaches, apart from the eased  
burden of carrying the papers. Besides, a few studies have reported negative results from digitization in data  
management for CHPs. Studies carried out by Zeleke et al. (2019) and Zaidi et al. (2020) highlighted that  
additional workload and security for mobile devices that capture and store data were a key concern among other  
factors for the users (Numair et al., 2021). They pointed out that a single accidental error in handling an electronic  
device can mess up a whole lot of data, and this could lead to a repeat in data collection (Chepkirui et al., 2025).  
Drawing from the TAM means that data handlers can stay away from mobile devices to avoid security risks for  
both the devices and themselves. The unfortunate incident of a system failure leading to data loss would result  
in an additional workload. This issue is already contentious, considering the many tasks a Community Health  
Promoter is expected to deliver in addition to data management.  
Using technology to collect, share, or report data can potentially minimize the risks of loss or data manipulation  
associated with the transportation and handling of paper-based data collection and reporting tools. This is  
because collection and transmission are real-time, and backups and safe retrieval are possible. According to  
Zeleke et al. (2019), the use of paper-and-pen methods for data collection can result in lost questionnaires due  
to misplacement and wear and tear, especially when these tools are in transit or in storage. However, despite the  
benefits of digitization, technical issues such as poor connectivity, unavailability of electronic devices, lack of  
proper training on their use, or lack of capacity to maintain the electronic devices could affect the quality of data  
if not well addressed. According to Kreienbrinck et al. (2025) and Zaid et al. (2020), more investment is required  
to address the behavioral, technical, and organizational challenges of user adoption. According to Tamrat et al.  
(2022), this requires strategies in change management and strengthening human resource capacity to adopt and  
sustain the advantages of digitization and its associated outcomes.  
The mixed results in the literature present a research gap that this study aims to address. Therefore, this study  
seeks to establish the potential effect of digitization on the relationship between DMPs and the sustainability of  
CHPs.  
Conceptual Framework  
The dependent variable in this study is the sustainability of CHPs, measured by the following indicators:  
continued health benefits, ongoing program activities, and enhanced local capacity. The independent variable is  
DMPs, with data collection, data protection, and data utilization as indicators, and the moderating variable:  
digitization. The indicators for digitization are availability, acceptability, and usability. The conceptual  
framework illustrates the hypothesized relationship between the variables under study. Figure 1 below shows  
the relationships among the variables in this study.  
Figure 1: Conceptual Framework  
Page 5903  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
RESEARCH METHODOLOGY  
Research Philosophy  
This study examined the potential influence of DMPs on the sustainability of CHPs and the moderating effect  
of digitisation and environmental factors on these two variables. This study was based on a positivist paradigm,  
which was the most suitable for this study, as it aimed to explain and predict the relationship between the study  
variables, grounded in theory. The paradigm provided hypotheses, which were tested by operationalising the  
variables and measures identified in the literature to answer the research questions (Jaja et al., 2022). Using a  
survey to collect quantitative, measurable, and observable data was supported by a positivist approach, in which  
the conclusions were independent of the researchers' views and the results were generalizable (Park et al., 2020).  
This theoretical view enabled the researcher to distance themselves from participants’ responses (Park et al.,  
2020). Additionally, an exploratory rather than confirmatory approach, anchored in the PLS-SEM, aligned well  
with the positivist perspective (Sabol et al., 2023). The findings of this study further inform and clarify existing  
theories to answer the question of CHP sustainability.  
Research Design  
This study adopted a descriptive cross-sectional survey design to determine whether a significant relationship  
exists among variables at a single point in time. The objective was to establish the moderating effect of  
digitalization on THE relationship between DMP and the sustainability of CHP. The advantage of using the  
research design is that it is simple, cost-effective, and has a minimal burden on participants (Taris et al., 2021),  
who are already thought to be overloaded. The cross-sectional survey design also suits the purpose of this study,  
as it is both explanatory and predictive, and the results can be generalised (Jaja et al., 2022). Using this design,  
the investigator explored relationships between variables without introducing an intervention or treatment. This  
study collected and analysed quantitative data to derive conclusions on the research question. Using a  
quantitative approach, the researcher examined how digitisation was likely to impact the relationship between  
DMPs and the sustainability of CHPs.  
Population of the Study  
The population of this study was the Community Health Programmes (CHPs) in Nairobi County, as established  
by the Kenya Ministry of Health through the Kenya Community Health Strategy (2018-2025). CHPs are an  
effective way to deliver healthcare and address the significant burden of disease globally (MoH, 2021). In the  
community health structure, CHPs are implemented through a Community Health Unit (CHU), a level-one  
health service and the smallest unit. According to the MOH (2021) strategy, a community health unit is a  
geographical area covering a population of about 5000 people, assigned to 10 Community Health Promoters and  
one supervisor (also called a Community Health Assistant). The community health units are connected to a  
health facility where they receive health services in addition to those provided by Community Health Promoters  
at the household or community level. In this study, the community health unit was the unit of analysis. In Nairobi  
County, there are 746 community health units distributed across the 17 Sub-Counties (MoH, 2021).  
Considering the time and resources available to conduct this study, a sample was selected to represent the 746  
community health units. The Community Health Promoters deliver all the preventive, promotional, and essential  
curative health services to the population at the community and household levels. They are also tasked with  
collecting, storing, and sharing data on these services and are expected to use the data whenever necessary to  
respond to emerging issues. Given the mentioned tasks, Community Health Promoters and their supervisors  
from the 17 Sub-Counties within Nairobi County were identified as the most suitable respondents for this study.  
Sampling Techniques  
The study population comprised 746 community health units. Since the population is large, a sample of the  
population was proposed and established through the PLS-SEM. The SEM model has various approaches for  
determining sample sizes. The use of PLS-SEM to determine sample size was preferred because it can achieve  
higher statistical power regardless of sample size, supports complex models, and accounts for prediction errors  
Page 5904  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
(Hair, 2022). PLS-SEM has been successfully used in other studies to determine sample size, such as Chirchir  
(2022) and Odock (2016). In line with SEM, the a-priori Sample Size Calculator for Structural Equation Models  
was used to compute the sample size based on the number of measurement indicators and latent variables (Soper,  
2020). The formula for calculating the sample size is as demonstrated below;  
The study involved 12 measurement indicators and 3 latent variables, yielding 15 observable variables. With a  
medium effect size of 0.3, a statistical power of 85%, and a significance level of 5%, the minimum required  
sample size for this study, according to Soper (2020), is 152 participants. A minimal non-response was expected  
in this study, as data collection was conducted in an in-person meeting; therefore, a 20% increase in the sample  
size is recommended (Bujang, 2021). Hence, the study's sample size was 190. This calculation is similar to that  
used successfully by other cross-sectional studies, such as Chirchir (2022) and Oredo (2016), which were  
deemed suitable for the conceptual model. A structured questionnaire with five sections was used to gather  
quantitative data for this study. The sample size was determined in proportion to the population strata, with the  
strata in this study being the 17 Sub-Counties in Nairobi County, Kenya.  
Diagnostic Tests  
Before analysis, data screening was performed to check for missing data, outliers, and multicollinearity. This  
included frequency, mean, and standard deviation. Then, the variance inflation factor (VIF) was used to test for  
multicollinearity and investigate the associations between study variables. There was no multicollinearity  
between the independent variables in this study as had been anticipated. Finally, heteroskedasticity was checked  
by analysing the residuals and plotting the residuals on a graph to show scatteredness. No heteroskedasticity was  
detected. A 5% significance level was used for the inferential statistics in this study.  
Data Analysis  
After the data screening process, which included checking for missing values, multicollinearity, scatter plots,  
and normality, the researcher carried out descriptive analysis to determine the demographic characteristics of the  
respondents and the characteristics of the community health units. This was done using SPSS software to  
measure the mean, standard deviation, frequencies and the mean of the study variables. Then, SmartPLS software  
from PLS-SEM was used to analyse data and establish the relationships among variables. The PLS-SEM was  
considered robust, with greater flexibility in analysis considering the following: it allows the modelling of  
multiple independent variables on the dependent variable, it can account for statistical errors, interactions, and  
correlations, and it can attain higher statistical power compared with Covariance-based Structural Equation (Hair  
& Alamer, 2022; Collier, 2020). PLS-SEM was also considered appropriate, due to the exploratory and  
predictive nature of this study (Hair & Alamer, 2022). PLS-SEM also allowed for simultaneous reliability,  
validity, and hypothesis testing (Shanmuganathan et al., 2022). PLS-SEM was deemed the most suitable for this  
study, where the sample size was 190, relatively small for covariance-based SEM. Chirchir successfully used  
the PLS-SEM statistical analysis technique (Chirchir, 2022). The measurement model was tested to specify the  
relationship among the variables and their observed indicators. The specification of the structural model and  
evaluation of the hypothesised relationship were established (Shanmuganathan et al., 2022) to answer the study  
question. Descriptive analysis will be used to analyse respondents’ demographic data using SPSS.  
Reliability and Validity Tests  
Reliability guarantees consistent and steady results from measuring a research phenomenon (Ahmed & Ishtiaq,  
2021). It is the degree to which results obtained by measurements and procedures can be replicated (Bahariniya  
et al., 2021). The researcher conducted a pilot study to gather valuable feedback on the tool and data collection  
process before commencing data collection for this study. Internal consistency was established by calculating a  
Cronbach's Alpha coefficient, which was considered a suitable measure of reliability for Likert scales, as Mohd  
and Lay (2021) argued. For this study, an internal consistency coefficient of .70 or higher was acceptable (Hair  
et al., 2022).  
Page 5905  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
Furthermore, principal component analysis was performed to assess the scale's reliability. The researcher aimed  
for a variance of at least 0.3 for any item considered part of the latent variable (Al-Emran et al., 2019). Composite  
reliability was used to evaluate the internal consistency of the latent constructs in the model (Al-Emran et al.,  
2019). Composite reliability was established if the score was more significant than 0.6. The researcher used the  
Average Variance Extracted (AVE) to establish the model's internal consistency. This study considered any  
AVE greater than 0.5 acceptable (Fauza, 2022; Al-Emran et al., 2019).  
The researcher further assessed the instruments' validity, focusing on content, construct, and criterion validity.  
Testing content validity ensured that all essential items were included in the study instrument. Experts in  
community health were consulted to validate the constructs and ensure that the items' interpretation and meaning  
were well understood. The instrument's criterion validity aligned with the literature and was reviewed by experts,  
after which it was adjusted in readiness for data collection.  
Two criteria were used to evaluate discriminant validity. First, the AVE for each construct was used to compare  
with its squared correlations with other constructs in the model. Discriminant validity was established if all AVE  
were higher than the corresponding inter-construct squared correlations (Fornell & Larcker, 1981). Heterotrait-  
Monotrait statistics were also assessed to establish the discriminant validity (Al-Emran et al., 2019).  
Discriminant validity was confirmed if HTMT ≤ 0.85 and its confidence interval excluded 1. The structural  
model was tested for collinearity among the constructs. Collinearity was established if the Variance Inflation  
Factor (VIF) is not more than 5.  
Structural Modelling Estimation and Hypothesis Testing  
Structural models were used to understand, predict, and establish the direction of the path, as the study’s aim  
was prediction and exploration. Assessment of the measurements and structural models enabled the researcher  
to determine the model’s predictive capabilities and examine how the constructs relate (Al-Emran et al., 2019).  
The model’s goodness of fit was established using a standard root mean square in line with the PLS-SEM analytic  
technique (Hair et al., 2022). If the difference between the observed and predicted results was zero, this was  
considered a perfect fit. However, the emphasis when establishing structural model solutions was more on  
construct reliability and validity, as recommended in the literature for PLS-SEM analyses (Hair et al., 2022;  
Shanmuganathan et al., 2022).  
The coefficient of determination (R2) was used to gauge the model’s predictive power. Also, marginal analysis  
was conducted to assess the effect of an omitted predictor on the R2 values for all dependent variables. In this  
case, R2 values of 0.19 indicated weak predictive power, 0.33 moderate predictive power, and 0.67 significant  
predictive power (Fauzi, 2022). Further, the effect of any omitted predictor on the value of R2 on the dependent  
variable (effect, f2) was established through marginal analysis. The f2 values of 0.35 signified large, while those  
of 0.15 and 0.02 indicated medium and small effects, respectively (Shanmuganathan et al., 2022). The researcher  
then evaluated the model’s predictive precision and relevance. Path model predictive precision is acceptable if  
Stone-Geisser’s Q2 > 0 (Sarstedt et al., 2021). In assessing the significance of path coefficients, a t-value> 1.96  
or a p-value < 0.05 at the 5% level of significance (2-tailed test) was considered substantial (Chirchir, 2022). In  
this case, the value zero was not included within the confidence interval.  
FINDINGS  
The targeted sample size was 190 participants, of whom 176 of the respondents replied. After cleaning the data,  
19 questionnaires were not fully answered and were removed from the analysis. Thus, 157 questionnaires were  
usable, representing a response rate of 83%, which is excellent for the study, as averred by Mugenda and  
Mugenda (2003).  
Sampling Adequacy and Sphericity Test  
This section presents sampling adequacy and sphericity tests to determine whether factor analysis is appropriate.  
Kaiser-Meyer-Olkin (KMO) metrics were employed to determine sample adequacy. Kaiser (1974) stated that  
KMO levels less than 0.5 are unacceptable. Dimension reduction is assessed using Bartlett's sphericity test. This  
Page 5906  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
is achievable with p-values < 0.05. All KMO measurements exceeded the specified minimum, with p values <  
0.05. This shows that all constructs are statistically significant. The results are shown in Table 1.  
Indicators  
KMO Value  
0.880  
Approx. Chi-Square  
703.624  
df Sig.  
Data Management Practices (DMP_1)  
10 0.000  
10 0.000  
10 0.000  
DMP_2  
DMP_3  
Digitization (D_1)  
D_2  
0.714  
346.033  
0.858  
527.679  
0.780  
252.972  
6
0.000  
0.851  
511.549  
10 0.000  
10 0.000  
10 0.000  
D_3  
0.850  
762.745  
Sustainability  
of  
Community  
Health 0.846  
310.816  
Programmes (SCHP_1)  
SCHP_2  
0.853  
0.867  
489.603  
648.751  
10 0.000  
10 0.000  
SCHP_3  
Table 1: KMO and Bartlett's Test Results  
Outer Model Indicator Reliability  
The indicator reliability for subconstructs of the three latent variables was evaluated. The results are presented  
in Table 2. All the indicators of the three latent constructs had outer factor loadings greater than the minimum  
required level of 0.5. Also, they were all statistically significant at the 5% level. In addition, the indicator  
reliability of all the indicators was above the threshold of 0.3. Hence, they were all retained for further analysis.  
Latent Variable  
DMP  
Indicator  
DMP_1  
DMP_2  
DMP_3  
D_1  
Outer Loading  
0.870  
Indicator Reliability  
0.757  
T statistic  
12.993  
26.051  
26.761  
5.677  
P value  
0.000  
0.000  
0.000  
0.000  
0.000  
0.000  
0.000  
0.000  
0.000  
0.883  
0.780  
0.890  
0.792  
D
0.574  
0.329  
0.889  
0.790  
29.728  
64.068  
40.556  
24.757  
34.122  
D_2  
0.921  
0.848  
D_3  
SCHP  
0.912  
0.832  
SCHP_1  
SCHP_2  
SCHP_3  
0.909  
0.826  
0.925  
0.856  
Table 2: Reflective Outer Model Results  
Internal Consistency Reliability  
Composite reliability and Cronbach’s Alpha values were used to determine the internal consistency reliability  
of the constructs. From the results in Table 3, the composite reliability values for all variables are above 0.8,  
Page 5907  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
which is higher than the minimum threshold of 0.7 (Hair et al., 2011). Also, the Cronbach’s Alpha for all  
variables is above the minimum required level of 0.7. Thus, internal consistency reliability is confirmed.  
Latent Construct Cronbach’s Alpha Composite Reliability AVE  
0.736  
0.856  
0.903  
0.867  
0.856  
0.905  
0.656  
0.776  
0.837  
D
DMP  
SCHP  
Table 3: Cronbach’s Alpha, Composite Reliability and AVE Results  
Convergent Validity  
Convergent validity was assessed using AVE and CFA. The results for CFA are presented in Table 4. The cross-  
loadings of subconstruct items to their corresponding latent constructs are higher than for other constructs. Also,  
the AVEs values in Table 3 for the latent variables are all above the minimum required value of 0.5. Therefore,  
convergent validity is confirmed (Hair et al., 2021).  
Indicator  
DMP_1  
DMP_2  
DMP_3  
D_1  
D
DMP  
0.870  
0.883  
0.890  
0.332  
0.535  
0.685  
0.623  
0.619  
0.667  
SCHP  
0.616  
0.606  
0.619  
0.356  
0.646  
0.857  
0.912  
0.909  
0.925  
0.575  
0.595  
0.602  
0.574  
0.889  
0.921  
0.722  
0.721  
0.792  
D_2  
D_3  
SCHP_1  
SCHP_2  
SCHP_3  
Table 4: Confirmatory Factor Analysis Statistics  
Discriminant Validity  
Discriminant validity was evaluated using two criteria: cross-loadings and the Heterotrait-Monotrait (HTMT)  
ratio (Henseler et al., 2014). From Table 4, all constructs load more heavily on their subconstructs than on any  
other. Additionally, the HTMT ratios in Table 5 were all below the maximum threshold of 0.9, except for SCHP  
<-> D, which was 0.939. However, discriminant validity is confirmed by the confirmatory factor analysis  
(Hensler et al., 2015; Hair et al., 2021; Rasoolimanesh, 2022).  
HTMT Ratios  
0.806  
0.939  
DMP <-> D  
SCHP <-> D  
Page 5908  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
0.791  
SCHP <-> DMP  
Table 5: HTMT ratios  
Collinearity for the outer model was determined using VIF and tolerance values. The results are presented in  
Table 6.  
Indicator  
DMP_1  
DMP_2  
DMP_3  
D_1  
Tolerance  
0.502  
0.455  
0.444  
0.844  
0.469  
0.481  
0.352  
0.360  
0.333  
VIF  
1.991  
2.198  
2.254  
1.185  
2.133  
2.081  
2.842  
2.777  
3.001  
D_2  
D_3  
SCHP_1  
SCHP_2  
SCHP_3  
Table 6: Variance Inflation Factor Outcomes  
It was noted that all the VIF values of the subconstructs were less than 5, while tolerance values were larger than  
the minimum of 0.2. Therefore, there is no multicollinearity in the outer model (Hair et al., 2021).  
The collinearity tests for the inner model are displayed in Table 7.  
Latent Construct  
Collinearity Statistics  
Tolerance  
0.395  
VIF  
2.529  
1.901  
2.126  
D -> SCHP  
0.526  
DMP -> SCHP  
D x DMP -> SCHP  
0.470  
Table 7: Collinearity and Tolerance Results  
It can be observed that all the VIF scores are below the maximum required level of 5 whereas all the tolerance  
values are greater than the threshold of 0.2. Therefore, there is no collinearity in the inner model.  
Predictive Relevance for Endogenous Variables and Overall Model Fit  
The inner model's predictive relevance for the endogenous variable SCHP was 0.740. This Q2 predictive value  
is significantly greater than the zero criterion. Predictive significance is therefore confirmed. Using the SRMR  
statistic, the overall goodness-of-fit of the moderation model was 0.074. This value is less than the required 0.1  
upper limit. Thus, it can be concluded that the model is fit.  
Page 5909  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
Target Endogenous Variable Variance  
The coefficient of determination, R2, for the endogenous variable SCHP, is displayed in Figure 2. The value is  
0.738. Thus, 73.8% of the variance in SCHP is associated with the variance in DMP and D. As per the study by  
Peng and Lai (2012), R2 values of 19%, 33%, and 67% represent low, moderate, and large explained variance,  
respectively. Based on these criteria, 73.8% is large. Hair et al. (2021) suggest that effect sizes of 0.025, 0.01,  
and 0.005 represent more reasonable standards for large, moderate, and low effect sizes, respectively. f2 for D,  
D as moderator, and DMP are 0.348, 0.125, and 0.097, respectively, and are all in the range of large effects as  
represented in Table 8.  
Path  
f-square Inference  
0.348  
0.125  
0.097  
Large Effect  
Large Effect  
Large Effect  
D -> SCHP  
D x DMP -> SCHP  
DMP -> SCHP  
Table 8: f2 Results  
Figure 2: R2 and f2 Statistics  
Moderation Analysis  
The moderating impact of D on the relationship between DMP and SCHP was done using the two-stage  
approach. Henseler and Chin (2010) proposed that when the main aim is to determine the significance of the  
moderation effect, the two-stage technique is appropriate, as it also results in a higher level of statistical power  
compared to the other methods (orthogonalizing and product indicator approaches).  
The PLS SEM moderating results are illustrated in Figure 3. The moderating effect has a value of -0.087, while  
the simple effect of DMP on SCHP is 0.236. These results suggest that the relationship between DMP and SCHP  
is 0.219 for an average level of D. Nevertheless, if D is increased by one unit, the relationship between DMP  
and SCHP will decrease by the interaction effect (that is, 0.219 + (– 0.087) = 0.132). In contrast, if D is reduced  
by one unit, the relationship between DMP and SCHP will increase by the interaction effect (i.e., 0.219 – (-  
0.087) = 0.306). The graphical presentation is presented in Figure 4.  
Figure 3: Structural Equation Model having R2 and Path Coefficients  
Page 5910  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
Figure 4: Simple Slope Plot for Moderating Effect  
The statistical significance of the moderating effect was tested. From the results in Table 9, Figure 5 and Figure  
6, it is observed that the moderating effect of digitization on the relationship between DMP and SCHP is  
statistically significant (t=2.070, p=0.038).  
Path Coeff. T Statistic P Statistic  
0.219  
2.959  
2.070  
0.003  
0.038  
DMP -> SCHP  
-0.087  
D x DMP -> SCHP  
Table 9: Moderating Effect Statistics  
Figure 5: Path Coefficients and P-values Moderation  
Figure 6: Path Coefficients and T-values Moderation  
Page 5911  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
DISCUSSION OF THE FINDINGS  
This study proposed a null hypothesis, stating that digitisation has no significant moderating effect on the  
relationship between data management practices and sustainability of community health programmes. Using  
Partial Least Squares Structural Equation Modeling, the analysis found a statistically significant moderate effect  
with a high model explanatory power and an acceptable model fit. These findings suggest that the impact of data  
management practices on the sustainability of community health programmes decreases with efforts on  
digitalization.  
This is based on the premise that digitization of data management practices, such as data collection, storage,  
retrieval, sharing, and use, makes it more efficient and improves the overall quality of data for timely and  
comprehensive decision-making, hence, improving sustainability of the community health programmes (Dillip  
et al., 2024; Kansiime et al., 2024). Consequently, when digitization is enhanced, the additional benefit of data  
management practices becomes insignificant, perhaps because digital systems already streamline data collection,  
storage, retrieval, sharing, and analysis for effective decision-making. This means that more efforts to enhance  
data management practices are needed in settings where digitization is unavailable or inadequate, to support the  
sustainability of community health programmes. It could also mean that with well-automated systems, data  
management practices are inevitably solved.  
The results of this study are consistent with recent studies by Demirci and Yardan (2023), who underscore the  
foundational role of data management practices in digital health environments, and David et al. (2024), who  
report that digital tools enhance community health initiatives, but with diminishing marginal returns in over-  
digitised environments. Similar context-sensitive effects of digitisation have been documented by Rîndașu et al.  
(2023) and Ribeiro-Navarrete et al. (2023), confirming that digitalization’s impact is contingent on  
complementary organisational capabilities. Blondino et al. (2024), Njororai et al. (2021), and Zaidi et al. (2020)  
found that digital tools increase data collection speed and accuracy, reduce manual workload, and enhance  
community health workers' capacity to deliver services effectively. These outcomes are achieved through direct  
digital entry, voice recordings, and image capture, all of which enable real-time data sharing and utilization.  
Jeilani and Hussein (2025) further assert that digitized data management eases administrative burden, which  
supports sustainability by reducing staff burnout and low morale.  
However, findings from Blondino et al. (2024), Numair et al. (2021), and Zaidi et al. (2020) also caution that  
without adequate training, supervision, support, and motivation to maintain high performance, the introduction  
of digital tools can initially increase workloads, negatively lower morale, and hinder the intended benefits. A  
different perspective is highlighted by Russpatrick et al. (2021), who point out that digitization is a barrier to  
performance and sustainability of programmes particularly if there is no proper integration and usability support.  
This argument is supported by the TAM, which proposes that the adoption of technology is based on perceived  
benefits and ease of use (Nyimbili & Chalwe, 2023). Hence, in this case, digitization will not be adopted if it is  
perceived to increase workload or interrupt the status quo.  
CONCLUSIONS OF THE RESEARCH  
This section presents the conclusions derived from this research, following the testing of the hypotheses designed  
to explore the variables of this study. Firstly, this research confirms that data management practices such as data  
collection, protection, and utilization influence the sustainability of community health programmes, resulting in  
enhanced local capacity and interventions that continue to benefit the host community (Tshuma et al., 2024;  
Habte et al., 2022; Ceptureanu & Ceptureanu, 2019). The results of this study highlight the critical role of data  
management practices in maintaining the continuity of community health programmes; if omitted, there would  
be a marked deterioration. It also confirms that the sustainability of community health programmes benefits from  
the routine data collected during implementation, thereby assisting with evidence-based planning and context-  
specific adaptations to emerging challenges, leading to programme ownership, improved health services, and  
sustainability (Tshuma et al., 2024; Hujala et al., 2021; Popa & Ștefan, 2019). Additionally, this study  
emphasised the tenacity of traditional, non-automated data management practices in contexts where digital  
infrastructure is absent or not adequately supported to function optimally, thereby facilitating the sustainability  
Page 5912  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
of community health programmes. This is especially beneficial in low-resourced settings where digitization  
adaptability is challenged by internet and electricity connectivity.  
Secondly, this study concludes that digitization had a significant influence on the relationship between data  
management practices and the sustainability of community health programmes. However, this influence has an  
inverse effect, meaning that increased digitisation reduces the impact of data management practices on the  
sustainability of community health programmes. Therefore, in settings where operations are highly automated,  
there is no need to invest in improving data management practices, perhaps because automation of programmes’  
operations fundamentally addresses challenges that may occur with manual programme administration,  
including data management (Owoyemi et al., 2022). Therefore, investments in such contexts should instead  
address challenges that could hinder the full implementation of digitization, as this has a direct impact on the  
sustainability of community health programmes, and override efforts to improve data management practices.  
Additionally, digital interventions in data management practices must be applied multidimensionally,  
recognising their interdependence and interconnectedness, as elaborated by the systems theory supporting this  
study (Kaboré et al., 2023; Donessouné et al., 2023). Therefore, the accessibility, availability, and usability of  
digital formats, combined with the requisite support such as training, supervision, feedback, and guidelines, are  
necessary to achieve the programme sustainability outcomes (Bogale et al., 2023).  
Implications of the Research  
Digital infrastructure availability in this study refers not only to access to mobile phones but also to the  
connectivity and power supply that accompany them. The TAM assumes adequate coverage of digital  
infrastructure to support data management practices for the sustainability of community health programmes;  
hence, it focuses on acceptance and use. Therefore, in a setting where, despite high mobile phone coverage,  
connectivity and power backup are inadequate, TAM does not predict adoption, which is critical to the  
sustainability of community health programmes. This study therefore proposes that perceived access to  
technology is an additional component of the Technological Accessibility and Acceptance Model, such that  
acceptability and adoption of technology are based on perceived benefit, ease of use, and ease of access.  
Additionally, this will cover not only the structural support for technology but also explain the cost of using  
digital infrastructure in low-resource settings, including a substantial segment of the population in this study.  
This way, both psychological preparedness, as a requisite of human behaviour, and the understanding of  
prevailing reality will be complemented by the best outcome.  
This study has made significant contributions to practice and policy in several ways. The outcome of this study  
is that there exists a strong relationship between data management practices and the sustainability of community  
health programmes. This relationship can be enhanced by digitisation and a supportive environment, which is  
realized through the implementation of policies and programmes. It points to where government leaders should  
invest in the design and planning of community health programmes to ensure their sustainability. The outcomes  
inform policymakers of the critical role of quality data in addressing implementation challenges that might hinder  
the continuity of health programmes.  
This study has further demonstrated the need to consider each and all aspects of data management subsets (data  
collection, data protection, and data utilisation) contributing to the whole system as supported by the system’s  
theory, hence highlighting the need to address all the aspects of data management to attain sustainability of  
community health programmes. Similarly, as supported by the social ecological theory, this study acknowledges  
the role of multi-layered interactions towards decision-making, and eventual behaviour among data handlers.  
According to this study, this could assist the practitioners and policymakers acknowledge and identify capacity  
needs (such as training and infrastructure provision) for different stakeholders at different levels of data  
management process. It highlights the aspects to be considered in planning for capacity enhancement to achieve  
the desired outcomes for the sustainability of community health programmes.  
RECOMMENDATIONS  
The study has established that, despite high phone coverage in Kenya and most Sub-Saharan African countries,  
the requisite support for connectivity and power to support functionality is lacking. Therefore, policymakers and  
Page 5913  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
funding agencies should prioritise investments in internet and electricity connectivity to support the existing and  
future efforts in digitalisation. Meanwhile, efforts to support traditional, non-digital data management practices  
(such as paper and pen) should continue, as this study has established that in settings where digitisation is not  
fully established, the influence of data management practices on the sustainability of community health  
programmes is high. This also means that in areas where digitisation is being introduced, programme  
implementors should focus their efforts on enhancing a hybrid system that combines capacity in the new  
technology with existing data management practices to ensure a seamless transition. Additionally, implementors  
should employ gradual digitisation, considering prevailing contextual factors such as connectivity to power and  
the internet, available resources, and the readiness of data handlers to digitise.  
This study used TAM to explain adaptive behaviour to technology. However, this model failed to explain the  
aspect of perceived accessibility as a key ingredient in its explanation. This study recommends the addition of  
this aspect to extend the TAM theory; perceived ease of access. Further research can be conducted on the  
proposed model to expand the body of knowledge. Similarly, the KMT failed to fully explain contextual aspects,  
considering that knowledge is expected to be from internal as well as external sources in different contexts. This  
study proposes extending the theory to a context-specific knowledge management framework, encompassing  
various contexts (for example, well-resourced or under-resourced, analogue or digital technology, formal or  
informal), to be applied in diverse settings. Future research could focus on the proposed extension of this theory  
to add to the body of knowledge.  
Shortcomings of the Study  
This study had a few limitations that could provide opportunities for further exploration by future researchers. It  
employed a descriptive cross-sectional survey design, utilising a quantitative approach. This data collection  
method lacks in-depth exploration, such as a qualitative approach, which may have limited valuable insights into  
explaining the research variables. Additionally, considering the limited time and resources available for  
conducting this study, the questionnaire was sent to participants via a link to their phones using Google Forms.  
Although the response rate was high, most did not attempt the open-ended questions that needed typing in digital  
form. Physical administration of the questionnaire may have provided more robust data, as there would be an  
opportunity to clarify any issues if need be. Incorporation of other data collection methods could be considered  
in future research to enhance robustness.  
The research also targeted community health promoters (as the primary data handlers) who self-reported on the  
questions under each variable. This may have led to self-reported bias. Additionally, owing to the cost, this study  
focused on the experiences of community health promoters (as primary data handlers in a community health  
programme), which could have left important aspects of data verification or insights from other data handlers at  
other levels in the system. Data verification at various levels of the data management system in the Ministry of  
Health is an area that future researchers could focus on.  
Proposed Areas for Further Research  
Data analysis for this study was conducted using PLS-SEM instead of other approaches, such as covariance-  
based SEM, as it is an exploratory and predictive approach. This also helped with the analysis of multiple  
constructs at the same time without the need to collapse them, including those of individual indicators from  
moderation variables. Future research could focus on other data analysis approaches. The use of a cross-sectional  
survey design in this study may have resulted in some information being omitted, as this approach does not  
permit in-depth exploration of the context and reasoning behind the responses provided. Therefore, future  
research could employ other research designs to triangulate for more comprehensive perspectives.  
This research focused on government-sponsored community health programs guided by the national community  
health strategy. Since other programmes are publicly or privately sponsored, although few, future studies could  
focus on them all and probably draw comparisons between them. Additionally, this study focused on data  
management practices from the perspective of community health promoters (community health primary data  
handlers) and their supervisors. Future studies could expand the study population to include all stakeholders who  
interact with data management systems, thereby obtaining a more comprehensive view of this data as it reaches  
Page 5914  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
different levels in the hierarchy of decision-making. Additionally, verifying the data submitted by community  
health promoters could provide valuable information on the accuracy, completeness, and timeliness of the data  
received at that level.  
Finally, this study has pointed out a need to extend the TAM to include perceived access to cover the issues of  
availability of digital infrastructure, including the phone, internet, and power connectivity. Additionally, this  
study highlighted the need to extend the KMT theory to make it context-specific. Further research can delve into  
the two theories to expand their scope and apply them to broader issues and multiple contexts.  
REFERENCES  
1. Åhlfeldt, E., Isaksson, D., & Winblad, U. (2023). Factors Explaining Program Sustainability: A Study of  
the Implementation of a Social Services Program in Sweden. Health & Social Care in the  
Community, 2023(1), 1458305.  
2. Ahmed, I., & Ishtiaq, S. (2021). Reliability and Validity: Importance in Medical Research. Methods,  
12(1), 2401-2406.  
3. Al-Emran, M., Mezhuyev, V., & Kamaludin, A. (2019). PLS-SEM in information systems research: a  
comprehensive methodological reference. In Proceedings of the International Conference on Advanced  
Intelligent Systems and Informatics 2018 4 (pp. 644-653). Springer International Publishing.  
4. Bagozzi, R. P. (2007). The legacy of the technology acceptance model and a proposal for a paradigm  
shift. Journal of the association for information systems, 8(4), 3.  
5. Bahariniya, S., Ezatiasar, M., & Madadizadeh, F. (2021). A Brief Review of the Types of Validity and  
Reliability of Scales in Medical Research. Journal of Community Health Research, 10(2), 100-102.  
6. Blondino, C. T., Knoepflmacher, A., Johnson, I., Fox, C., & Friedman, L. (2024). The use and potential  
impact of digital health tools at the community level: results from a multi-country survey of community  
health workers. BMC Public Health, 24(1), 650.  
7. Bogale, T. N., Teklehaimanot, S. M., Fufa Debela, T., Enyew, D. B., Nigusse Bedada, A., Dufera  
Kebebew, S., ... & Bekele, T. A. (2023). Barriers, facilitators and motivators of electronic community  
health information system use among health workers in Ethiopia. Frontiers in Digital Health, 5, 1162239.  
8. Bujang, M. A. (2021). A step-by-step process on sample size determination for medical research. The  
Malaysian journal of medical sciences: MJMS, 28(2), 15.  
9. Ceptureanu, S. I., & Ceptureanu, E. G. (2019). Community-based healthcare programs sustainability  
impact on the sustainability of host organizations: A structural equation modeling analysis. International  
Journal of Environmental Research and Public Health, 16(20), 4035.  
10. Chepkirui, M., Dellicour, S., Musuva, R., Odero, I., Omondi, B., Omondi, B., ... & Taegtmeyer, M.  
(2025). Missed opportunities for digital health data use in healthcare decision-making: A cross-sectional  
digital health landscape assessment in Homa Bay County, Kenya. PLOS Digital Health, 4(6), e0000870.  
11. Chirambo, G. B., Hardy, V., Heavin, C., O’Connor, Y., O’Donoghue, J., Mastellos, N., ... & Thompson,  
M. (2018). Perceptions of a mobile health intervention for Community Case Management in Malawi:  
Opportunities and challenges for Health Surveillance Assistants in a community setting. Malawi Medical  
Journal, 30(1), 6-12.  
12. Chirchir, M. K. (2022). Supply Chain Integration, Competitive Advantage, Environmental Dynamism  
and Performance of Large-Scale Manufacturing Firms in Kenya (Doctoral dissertation, University of  
Nairobi).  
13. Collier, J. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques.  
Routledge.  
14. Davis, F. D. (1989). Technology acceptance model: TAM. Al-Suqri, MN, Al-Aufi, AS: Information  
Seeking Behavior and Technology Adoption, 205-219.  
15. Demirci, H. F., & Yardan, E. D. (2023). Data management in the digital health environment scale  
development study. BMC Health Services Research, 23(1), 1249.  
16. Dillip, A., Kahamba, G., Sambaiga, R., Shekalaghe, E., Kapologwe, N., Kitali, E., ... & Sarkar, N. (2024).  
Using digital technology as a platform to strengthen the continuum of care at community level for  
maternal, child and adolescent health in Tanzania: introducing the Afya-Tek program. BMC Health  
Services Research, 24(1), 865.  
Page 5915  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
17. Donessouné, F. M., Sossa, G. O., & Kouanda, S. (2023). Sustainability of community health programme  
using community-based organizations: a challenge for stakeholders. BMC Health Services Research,  
23(1), 1-11.  
18. Fauzi, M. A. (2022). Partial Least Square Structural Equation Modelling (PLS-SEM) in Knowledge  
Management Studies: Knowledge Sharing in Virtual Communities. Knowledge Management & E-  
Learning, 14(1), 103-124.  
19. Feroz, A., Jabeen, R., & Saleem, S. (2020). Using mobile phones to improve community health workers  
performance in low-and-middle-income countries. BMC public health, 20, 1-6.  
20. Fitzpatrick, K. M., Ody, M., Goveas, D., Montesanti, S., Campbell, P., MacDonald, K., ... & Roach, P.  
(2023). Understanding virtual primary healthcare with Indigenous populations: a rapid evidence review.  
BMC Health Services Research, 23(1), 303.  
21. Fornell, C., & Larcker, D. F. (1981). Structural equation models with unobservable variables and  
measurement error: Algebra and statistics: 382-388  
22. Habte, D., Zemenfeskudus, S., Endale, M., Zeidan, M., Getachew, D., Woldemichael, D., ... & Abayneh,  
S. A. (2022). Enhancing and promoting data management and systematic monitoring for an improved  
HIV/AIDS programs in Addis Ababa, Ethiopia. BMC Health Services Research, 22, 1-11.  
23. Hair, J., & Alamer, A. (2022). Partial Least Squares Structural Equation Modeling (PLS-SEM) in second  
language and education research: Guidelines using an applied example. Research Methods in Applied  
Linguistics, 1(3), 100027.  
24. Hujala, T., & Laihonen, H. (2021). Effects of knowledge management on the management of health and  
social care: a systematic literature review. Journal of knowledge management, 25(11), 203-221.  
25. Hussein, S., Otiso, L., Kimani, M., Olago, A., Wanyungu, J., Kavoo, D., ... & Karuga, R. (2021).  
Institutionalizing community health services in Kenya: A policy and practice journey. Global Health:  
Science and Practice, 9(Supplement 1), S25-S31.  
26. Jaja, I. R., Idoniboye, O., & Amadi, C. C. (2022). A Critique of the Positivist Paradigm in Human  
Sciences. International Journal of Research and Innovation in Social Science, 6(3), 224-229.  
27. Jeilani, A., & Hussein, A. (2025). Impact of digital health technologies adoption on healthcare workers’  
performance and workload: perspective with DOI and TOE models. BMC Health Services Research,  
25(1), 271.  
28. Kaboré, S. S., Ngangue, P., Soubeiga, D., Barro, A., Pilabré, A. H., Bationo, N., ... & Savadogo, G. B.  
L. (2022). Barriers and facilitators for the sustainability of digital health interventions in low and middle-  
income countries: a systematic review. Frontiers in digital health, 4, 1014375.  
29. Kansiime, W. K., Atusingwize, E., Ndejjo, R., Balinda, E., Ntanda, M., Mugambe, R. K., & Musoke, D.  
(2024). Barriers and benefits of mHealth for community health workers in integrated community case  
management of childhood diseases in Banda Parish, Kampala, Uganda: a cross-sectional study. BMC  
Primary Care, 25(1), 173.  
30. Khatri, R. B., Endalamaw, A., Erku, D., Wolka, E., Nigatu, F., Zewdie, A., & Assefa, Y. (2024). Enablers  
and barriers of community health programs for improved equity and universal coverage of primary health  
care services: A scoping review. BMC primary care, 25(1), 385.  
31. Kirk, K., McClair, T. L., Dakouo, S. P., Abuya, T., & Sripad, P. (2021). Introduction of digital reporting  
platform to integrate community-level data into health information systems is feasible and acceptable  
among various community health stakeholders: A mixed-methods pilot study in Mopti, Mali. Journal of  
Global Health, 11.  
32. Lee, J., Lynch, C. A., Hashiguchi, L. O., Snow, R. W., Herz, N. D., Webster, J., ... & Erondu, N. A.  
(2021). Interventions to improve district-level routine health data in low-income and middle-income  
countries: a systematic review. BMJ global health, 6(6).  
33. Mehra, R., Boyd, L. M., Lewis, J. B., & Cunningham, S. D. (2020). Considerations for building  
sustainable community health worker programs to improve maternal health. Journal of Primary Care &  
Community Health, 11, 2150132720953673.  
34. Ministry of Health, Division of Community Health Services, Republic of Kenya (2021). Kenya  
Community Health Policy 2020 – 2030. Nairobi, Kenya. https://www.health.go.ke/  
35. Ministry of Health, Division of Community Health Services, Republic of Kenya (2019). Kenya  
Community Health Strategy 2020 – 2025. Nairobi, Kenya. https://www.health.go.ke/  
Page 5916  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
36. Ministry of Health, Division of Community Health Services, Republic of Kenya (2021). National  
Community Health Digitization Strategy 2020-2025. Nairobi, Kenya. https://www.health.go.ke/  
37. Ministry of Health. (2020). Kenya Community Health Policy 2020–2030. Nairobi.  
38. Mohd Dzin, N. H., & Lay, Y. F. (2021). Validity and reliability of adapted self-efficacy scales in  
Malaysian context using PLS-SEM approach. Education Sciences, 11(11), 676.  
39. Mremi, I. R., Sindato, C., Kishamawe, C., Rumisha, S. F., Kimera, S. I., & Mboera, L. E. (2022).  
Improving disease surveillance data analysis, interpretation, and use at the district level in Tanzania.  
Global Health Action, 15(1), 2090100.  
40. Muinga, N., Magare, S., Monda, J., English, M., Fraser, H., Powell, J., & Paton, C. (2020). Digital health  
Systems in Kenyan Public Hospitals: a mixed-methods survey. BMC medical informatics and decision  
making, 20(1), 2.  
41. Njororai, F., Ganu, D., Nyaranga, K. C., & Wilberforce, C. (2021). Role of socio-demographic and  
environmental determinants on performance of community health Workers in Western  
Kenya. International Journal of Environmental Research and Public Health, 18(21), 11707.  
42. Numair, T., Harrell, D. T., Huy, N. T., Nishimoto, F., Muthiani, Y., Nzou, S. M., ... & Kaneko, S. (2021).  
Barriers to the digitization of health information: a qualitative and quantitative study in Kenya and Lao  
PDR using a cloud-based maternal and child registration system. International Journal of Environmental  
Research and Public Health, 18(12), 6196.  
43. Nyimbili, L., & Chalwe, M. (2023). A Review of Technology Acceptance and Adoption Models and  
Theories. International Journal for Multidisciplinary Research (IJFMR), 5(6), 1-10.  
44. Odock, S. O. (2016). Green supply chain management practices and performance of ISO 14001 certified  
manufacturing firms in East Africa (Doctoral dissertation, University of Nairobi).  
45. Ogutu, M., Muraya, K., Mockler, D., & Darker, C. (2021). Factors influencing the performance of  
community health volunteers working within urban informal settlements in low-and middle-income  
countries: a qualitative meta-synthesis review. Human resources for health, 19, 1-21.  
46. Owoyemi, A., Osuchukwu, J. I., Azubuike, C., Ikpe, R. K., Nwachukwu, B. C., Akinde, C. B., ... &  
Olaniran, S. (2022). Digital solutions for community and primary health workers: lessons from  
implementations in Africa. Frontiers in Digital Health, 4, 876957.  
47. Park, Y. S., Konge, L., & Artino Jr, A. R. (2020). The positivism paradigm of research. Academic  
medicine, 95(5), 690-694.  
48. Perry, H. B., Chowdhury, M., Were, M., LeBan, K., Crigler, L., Lewin, S., ... & Hodgins, S. (2021).  
Community health workers at the dawn of a new era: 11. CHWs leading the way to “Health for All”.  
Health Research Policy and Systems, 19, 1-21.  
49. Popa, I., & Ștefan, S. C. (2019). Modeling the pathways of knowledge management towards social and  
economic outcomes of health organizations. International journal of environmental research and public  
health, 16(7), 1114.  
50. Ramli, N. A., Latan, H., & Nartea, G. V. (2018). Why should PLS-SEM be used rather than regression?  
Evidence from the capital structure perspective. Partial least squares structural equation modeling:  
Recent advances in banking and finance, 171-209.  
51. Regeru, R. N., Chikaphupha, K., Bruce Kumar, M., Otiso, L., & Taegtmeyer, M. (2020). ‘Do you trust  
those data?’—a mixed-methods study assessing the quality of data reported by community health  
workers in Kenya and Malawi. Health Policy and Planning, 35(3), 334-345.  
52. Rîndaşu, S. M., Ionescu-Feleagă, L., Ionescu, B. Ş., & Topor, I. D. (2023). Digitalisation and skills  
adequacy as determinants of innovation for sustainable development in EU countries: a PLS-SEM  
approach. Amfiteatru Economic, 25, 968-986.  
53. Russpatrick, S., Li, M., Braa, J., Bruland, A., Rodvelt, M. O., Muhire, A., ... & Rustad, S. (2021).  
Improving Data Use and Participatory Action and Design to Support Data Use: The Case of DHIS2 in  
Rwanda. arXiv preprint arXiv:2108.09721.  
54. Saade, R., Nebebe, F., Mak, T., & Leung, N. K. (2011). Knowledge management systems development:  
Theory and practice. Interdisciplinary Journal of Information, Knowledge & Management, 6.  
55. Sabol, M., Hair, J., Cepeda, G., Roldán, J. L., & Chong, A. Y. L. (2023). PLS-SEM in information  
systems: seizing the opportunity and marching ahead full speed to adopt methodological updates.  
Industrial Management & Data Systems, 123(12), 2997-3017.  
Page 5917  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)  
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025  
56. Sarstedt, M., Ringle, C. M., & Hair, J. F. (2021). Partial least squares structural equation modeling. In  
Handbook of market research (pp. 587-632). Cham: Springer International Publishing.  
57. Shanmuganathan, S., Mustapha, F. I., & Wilson, A. (2022). Evaluating the sustainability of non-  
communicable diseases programs in Malaysia. BMC Public Health, 22(1), 1463.  
58. Silva, P. (2015). Davis' technology acceptance model (TAM)(1989). Information seeking behavior and  
technology adoption: Theories and trends, 205-219.  
59. Soper, D. S. (2020). A-priori sample size calculator for structural equation models [Software].  
https://www.danielsoper.com/statcalc/calculator.aspx?id=89  
60. Tamrat, T., Chandir, S., Alland, K., Pedrana, A., Shah, M. T., Footitt, C., ... & Mehl, G. L. (2022).  
Digitalization of routine health information systems: Bangladesh, Indonesia, Pakistan. Bulletin of the  
World Health Organization, 100(10), 590.  
61. Taris, T. W., Kessler, S. R., & Kelloway, E. K. (2021). Strategies addressing the limitations of cross-  
sectional designs in occupational health psychology: What they are good for (and what not). Work &  
Stress, 35(1), 1-5.  
62. Tshuma, N., Elakpa, D. N., Moyo, C., Soboyisi, M., Moyo, S., Mpofu, S., ... & Mtapuri, O. (2024). The  
transformative impact of community-led monitoring in the South African health system: a  
comprehensive analysis. International Journal of Public Health, 69, 1606591.  
63. Unkels, R., Manzi, F., Kapologwe, N. A., Baker, U., Ahmad, A., Nabiev, R., ... & Hirose, A. (2023).  
Feasibility, usability and acceptability of a novel digital hybrid-system for reporting of routine maternal  
health information in Southern Tanzania: A mixed-methods study. PLOS Global Public Health, 3(1),  
e0000972.  
64. World Health Organization (2020). Health policy and system support to optimize community health  
worker programmes for HIV, TB and malaria services: an evidence guide. Geneva: Licence: CC BY-  
NC-SA  
3.0  
IGO.  
https://iris.who.int/bitstream/handle/10665/340078/9789240018082-  
eng.pdf?sequence=1  
65. Yang, J. E., Lassala, D., Liu, J. X., Whidden, C., Holeman, I., Keita, Y., ... & Johnson, A. D. (2021).  
Effect of mobile application user interface improvements on minimum expected home visit coverage by  
community health workers in Mali: a randomised controlled trial. BMJ global health, 6(11),  
66. Zaidi, S., Kazi, A. M., Riaz, A., Ali, A., Najmi, R., Jabeen, R., ... & Sayani, S. (2020). Operability,  
usefulness, and task-technology fit of an mhealth app for delivering primary health care services by  
community health workers in underserved areas of Pakistan and Afghanistan: Qualitative study. Journal  
of Medical Internet Research, 22(9), e18414.  
67.  
Zeleke, A. A., Worku, A. G., Demissie, A., Otto-Sobotka, F., Wilken, M., Lipprandt, M., ... & Röhrig,  
R. (2019). Evaluation of electronic and paper-pen data capturing tools for data quality in a public health  
survey in a health and demographic surveillance site, Ethiopia: randomized controlled crossover health  
care information technology evaluation. JMIR mHealth and uHealth, 7(2), e10995.  
Page 5918  
www.rsisinternational.org