INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2624
www.rsisinternational.org
Use of Routine Health Data by County Health Management Teams in
Kenya: Evidence from a Quasi-experimental Study
Joshua Gitonga M'imaita, Prof. John Paul Oyore, Prof. George Ochieng Otieno
Kenyatta University, Kenya
DOI: https://dx.doi.org/10.47772/IJRISS.2025.91100210
Received: 10 November 2025; Accepted: 20 November 2025; Published: 05 December 2025
ABSTRACT
Strengthening the use of routine health information is an essential element of effective health system
governance in Kenya. Although the country has invested substantially in digital reporting platforms such as
District Health Information System 2 (DHIS2), the integration of Routine Health Data into managerial
planning and monitoring activities remains inconsistent across counties. County Health Management Teams
(CHMTs) often meet reporting obligations but do not always translate the available information into evidence
informed programme adjustments or performance review processes. This study examined the extent to which
CHMTs use routine data in their decision making and assessed whether a structured capacity building
programme could strengthen data use practices at the subnational level. A quasi-experimental approach
supported by a Difference in Differences model was applied. Twelve counties were selected from the national
regional economic blocs, with two chosen from each and allocated to either the intervention or control arm.
Baseline data were collected from two hundred CHMT members in August 2024, followed by a nine-month
intervention that involved targeted training sessions and continuous technical support. Endline data collection
took place in April 2025. A key limitation of the study was the reliance on self-reported measures which may
have been influenced by social desirability tendencies. The analysis produced a statistically significant
treatment effect with a coefficient of 0.4593 and a p value of 0.0046. CHMTs in the intervention counties
demonstrated notable improvements in incorporating routine data into programme monitoring policy
development and adjustment of ongoing interventions. These shifts were not observed in the control counties.
Further, subgroup analyses showed no statistically significant differences across gender, education level, age,
or duration of CHMT membership. The findings indicate that structured capacity building initiatives can
meaningfully strengthen routine data use among CHMTs. Sustaining these gains will require continued
investment in analytical competencies and the institutionalization of routine evidence review processes within
the devolved health sector.
Keywords: Routine health data; evidence-based decision-making; capacity-building; public health
governance; Kenya; health information systems.
INTRODUCTION
A. Background to the Study
The use of routine health data for decision-making has become a cornerstone of health system strengthening
globally. According to the World Health Organization’s 2017 Framework and Standards for Country Health
Information Systems, well-functioning health information systems are foundational to evidence informed
planning, priority setting, and the systematic monitoring of health sector performance. Countries that
effectively use routine data are better positioned to improve service delivery, allocate resources efficiently, and
achieve universal health coverage targets.
Despite global acknowledgement of the value of routine health information, many health systems continue to
struggle with transforming routinely collected data into meaningful inputs for management and planning. In
several low- and middle-income settings, routine data systems generate substantial volumes of information, yet
limited analytical capacity and weak data use cultures restrict their contribution to performance improvement
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2625
www.rsisinternational.org
(Nutley and Reynolds, 2013; AbouZahr and Boerma, 2015). These constraints highlight the need to strengthen
the institutional and technical conditions that enable routine data to inform health sector decisions.
Within sub-Saharan Africa, health information systems have expanded rapidly over the past two decades.
Many countries have adopted digital reporting platforms such as the District Health Information Software 2
(DHIS2), improving the completeness and timeliness of health data reporting. Despite these advancements,
evidence from across the region shows that health managers still underuse available data for operational and
strategic decisions. Studies in Zambia, Tanzania, and Uganda have documented similar trends where data
generated at facility and district levels are primarily used for reporting rather than for guiding planning,
budgeting, or programme improvement (Mutale et al., 2013; Ndabarora et al., 2014). Barriers include
inadequate data literacy, weak institutional support, limited human resource capacity, and a persistent culture
of compliance reporting rather than analytical interpretation (Oware et al., 2025; Zeng et al., 2022). These
challenges reveal that although data infrastructure has expanded, the behavioral and organizational conditions
required to embed data use within core CHMT functions such as planning, priority setting, supervision, and
performance review remain insufficiently developed.
The devolution of health services in 2013 fundamentally restructured Kenya’s health governance arrangements
by transferring considerable managerial authority to county governments (Ministry of Health, 2014). To
support this shift, the Ministry of Health implemented the Kenya Health Information System, a DHIS2 based
platform intended to standardize the collection, reporting, and analysis of routine health data nationwide
(Ministry of Health, 2018). KHIS was designed to enhance evidence informed planning at both national and
county levels, with County Health Management Teams expected to interpret and apply system generated data
to guide resource allocation, human resource deployment, programme monitoring, and service delivery
improvements (Oluoch et al., 2020).
Despite substantial gains in reporting completeness and digitalization through KHIS, persistent gaps remain in
the actual use of routine data for managerial and policy decisions. National assessments have documented
limited analytical capacity among county managers, weak feedback loops, and insufficient integration of data
analysis within planning and budgeting processes (Ministry of Health, 2018). Empirical studies similarly note
that many counties rely on informal practices and precedent rather than systematic review of routine health
information when making operational decisions (Oluoch et al., 2020). These constraints have limited the ability
of KHIS to function as an effective decision support tool, underscoring the continued need to strengthen data
use culture within CHMTs to enhance efficiency, accountability, and overall health system performance.
B. Problem Statement
The ability of health systems to make timely, evidence-informed decisions depends on the effective generation,
analysis, and utilization of routine health data. Despite global recognition of the importance of routine health
information systems (RHIS) in improving health system performance, many developing countries continue to
experience a persistent disconnect between data availability and its practical use in decision-making (Nutley &
Reynolds, 2013; AbouZahr & Boerma, 2015). In sub-Saharan Africa, national investments in digital reporting
platforms such as the District Health Information Software (DHIS2) have enhanced data reporting rates but
have not translated into consistent evidence-based decision-making (Mutale et al., 2013; Ndabarora et al.,
2014). Inadequate analytical capacity, poor data quality, and limited accountability mechanisms continue to
undermine the transformation of data into actionable intelligence.
As a result, health managers frequently generate substantial volumes of routine data that remain underutilized,
a pattern confirmed in recent Kenyan studies. Njuguna, Muiruri and Njoroge (2022) found that although public
health facilities in Nairobi County routinely collected service data, limited analytical capacity and inconsistent
participation in data review forums significantly constrained their use for planning and performance
management. Likewise, Oware et al. (2025) reported that across fifteen counties, DHIS2 data were widely
perceived as available but were often not applied in decision making due to inadequate analytic skills, low
prioritization of data use, and reliance on a small group of technical officers rather than broader managerial
engagement. These findings demonstrate that even where digital reporting systems are functional, routine data
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2626
www.rsisinternational.org
continue to have limited influence on resource allocation and programme adjustments within Kenya’s health
system. This underutilization represents a major bottleneck in the achievement of equitable, efficient, and
accountable health systems.
Kenya mirrors this regional challenge despite significant progress in data collection and digitization through
the District Health Information System 2 (DHIS2). The devolution of health services in 2013 transferred
substantial decision-making authority to county governments, thereby heightening the need for strong data-use
capacity among CHMTs. While reporting completeness within KHIS is relatively high in most counties, the
system’s data are still rarely used to guide operational or strategic decisions. Evidence from the Ministry of
Health’s National and County Health Budget Analysis FY 2018–2019 (Ministry of Health, 2019) and the
Kenya Health Financing System Assessment (Dutta, Maina, Ginivan & Koseki, 2018) shows that many
counties continue to base planning, budgeting, and staffing decisions on historical expenditure patterns, donor
driven priorities, and political considerations rather than on systematic analysis of routine health data. This
pattern is further reinforced by recent county level studies in Kitui and Marsabit, which found that managerial
decisions were often influenced more by precedent and external pressures than by empirical evidence (Karijo,
2021; Aila, 2021).
Weak feedback systems limited analytical skills, and minimal institutional incentives for data use exacerbate
this gap (Wako et al., 2018; O’Meara et al., 2022). This underuse of available data results in inefficient
resource allocation, misaligned priorities, and reduced accountability within Kenya’s devolved health system.
Despite multiple policy frameworks advocating for evidence-based management, limited empirical research
has systematically examined the types of decisions informed by routine health data at county level. This body
of evidence highlights that existing literature, while documenting barriers and describing the limited
application of RHD for planning and resource allocation (Masaviru, Namusonge, & Nambuswa, 2021;
Muwonge et al., 2022), lacks causal evaluation of interventions designed to enhance the organizational and
behavioural components of data use in devolved health systems (Wako et al., 2018; Nutley & Reynolds, 2013).
This knowledge gap constrains efforts to strengthen evidence based public health management, underscoring
the necessity of this study.
C. Objective of the Study
The overall objective of this study was to identify the most common public health decisions informed by
routine health data at the county level in Kenya and to examine how capacity-building interventions influenced
data-driven decision-making among CHMTs.
The study was guided by the following specific objectives:
1. To examine differences in the use of routine health data for public health decision-making between
baseline and endline across intervention and control counties.
2. To assess variations in the use of routine health data for decision-making across selected demographic
and institutional characteristics.
3. To evaluate the effect of a targeted capacity-building intervention on the integration of routine health
data into county-level public health decision-making.
The study also sought to assess the following hypotheses:
1. H₀₁: There is no statistically significant difference in the use of routine health data for public health
decision-making between intervention and control counties.
2. H₀₂: There are no statistically significant variations in the use of routine health data for decision-
making across demographic and institutional characteristics.
3. H₀₃: The training intervention has no statistically significant effect on the integration of routine health
data into public health decision-making among CHMTs.
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2627
www.rsisinternational.org
D. Significance of the Study
This study holds important implications for policy and practice within Kenya’s devolved health system. By
identifying the specific public health decisions that draw most heavily on routine health data, it clarifies how
evidence is actually applied within county governance structures. These insights help align future investments
in health information systems with the decision areas of greatest managerial relevance. The study also
contributes methodologically by demonstrating that structured capacity building can enhance the integration of
routine data into decision making. This addresses a key gap in understanding how routine data are translated
into actionable intelligence within everyday public health management. The findings are therefore valuable to
policymakers, development partners, and county health managers seeking to institutionalize evidence informed
governance and strengthen accountability in planning and resource allocation.
MATERIALS AND METHODS
A. Study Design and Setting
This study employed a quasi-experimental design with intervention and control groups. All forty seven
counties in Kenya formed the initial sampling frame for the study. To ensure balanced representation of the
country’s major geographic and administrative contexts, the counties were first grouped into the six officially
recognized regional economic blocs. From each bloc, two counties were randomly selected, resulting in a total
of twelve counties. This approach provided a stratified structure to the sampling process, with stratification
occurring at the bloc level and random selection occurring within each stratum. The final sample therefore
reflected both geographic diversity and the organizational variation present across Kenya’s devolved health
system.
Six counties formed the intervention arm and received a structured capacity building programme grounded in a
customized MEASURE Evaluation curriculum. The training was delivered over a nine-month period and
comprised modular sessions on data quality assessment, indicator interpretation, basic statistical analysis,
development of data review products, and application of routine data to planning and performance
management. These modules were implemented through a blended approach that combined in person
workshops, virtual coaching sessions, and on-site mentorship for CHMT members. The remaining six counties
served as the control arm and did not receive any training during the study period. Baseline and endline data
were collected from both groups to assess changes over time and to estimate the effect of the intervention on
routine data use.
From the Lake Region Economic Bloc, Kisii served as the intervention county and Kisumu as the control. In
the North Rift Economic Bloc, Uasin Gishu represented the intervention group and Turkana the control.
Kiambu and Meru were drawn from the Mt. Kenya and Aberdare Bloc, Garissa and Marsabit from the Frontier
Counties Development Council, Kilifi and Kwale from the JumuiyayaKaunti za Pwani, and Machakos and
Makueni from the Southeastern Kenya Economic Bloc. This distribution provided a balanced cross-section of
counties varying in resource capacity, infrastructure, and information system maturity.
The study was conducted within Kenya’s devolved health system, where CHMTs are responsible for
coordination, planning, and oversight of health service delivery (Government of Kenya, 2015; Muinga, Paton,
& English, 2018). Each CHMT served as a unit of analysis since it functions as the central mechanism through
which routine data are synthesized and translated into actionable decisions. Ethical approval was obtained from
a recognized Institutional Review Board, and official authorization was granted by the Ministry of Health. All
participants provided informed consent, and confidentiality was observed throughout the study.
B. Study Population and Sampling
The study targeted County Health Management Team members occupying administrative and technical
positions responsible for health sector management in the selected counties. This multidisciplinary group
comprised nurses, clinicians, health records officers, public health officers, and other senior administrators
who collectively oversee planning, resource allocation, supervision, and monitoring of county health systems
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2628
www.rsisinternational.org
(Government of Kenya, 2015; Muinga, Paton, & English, 2018). Sampling employed a combination of
purposive and simple random techniques. The six regional economic blocs in Kenya were first purposively
selected to ensure representation across diverse administrative and geographic contexts. Two counties were
then randomly selected from each bloc, yielding twelve counties that were equally assigned to the intervention
and control groups. All consenting CHMT members within the selected counties were enrolled in the study,
generating a total sample of 200 participants. The same individuals were followed and surveyed at both
baseline and endline, ensuring consistency in respondent identity and supporting the validity of the pre post
and Difference in Differences analyses.
The sample size for this study was determined using Cochran’s (1977) formula for proportions and
incorporated explicit assumptions to ensure adequate statistical power. A standard normal value of Z = 1.96
was used to represent a 95 percent confidence level. Because the true proportion of CHMT members routinely
using health data was unknown, an expected proportion of p = 0.50 was applied, which yields the maximum
variance and therefore the largest possible sample size. The desired margin of error was set at 0.10, allowing
the estimated proportion to vary within ten percentage points in either direction. Based on these assumptions,
the minimum required sample size was calculated to be 192 CHMT members. To accommodate potential
attrition, the target was increased to 200 participants, which constituted the final sample enrolled in the study.
C. Data Collection Procedures
Data were collected at two distinct time points, baseline and endline, to examine changes in the use of routine
health data for decision-making among CHMTs. The baseline phase established the pre-intervention conditions
for both the intervention and control counties, while the endline phase captured post-intervention outcomes
following the implementation of the data-use capacity-building program. A structured selfadministered
questionnaire was used as the main data collection instrument.
The data collection tool was developed in alignment with the study objectives and adapted from validated
instruments commonly used in assessments of routine health information system performance. Core sections
were drawn from the Performance of Routine Information System Management (PRISM) Diagnostic Tools
developed by MEASURE Evaluation, which assess technical, organizational and behavioral determinants of
routine data use. Additional items measuring decision making practices and data use frequency were also
adapted.
Data collection was carried out by trained research assistants under the direct supervision of the principal
investigator to maintain consistency and accuracy. Enumerators underwent comprehensive training on research
ethics, confidentiality, and standardized administration of the questionnaire. Although the capacity building
activities were delivered through scheduled training sessions, ongoing technical support, and structured
learning materials over a nine-month, the use of self-reported data still presents the possibility of social
desirability bias. Participants may have been inclined to portray their data use practices more favorably
because they were aware of the intervention’s objectives and had sustained exposure to training content.
D. Data Analysis Techniques
Data were analyzed using both descriptive and inferential statistical methods to address the study objectives.
Descriptive statistics, including frequencies, percentages, means, and standard deviations, were first computed
to summarize respondent characteristics and provide an overview of baseline and endline patterns in the use of
routine health data for decision-making. These descriptive analyses offered insights into how CHMTs applied
data across key functional areas such as planning, budgeting, monitoring, and resource allocation.
Inferential analysis was performed to determine whether observed differences in data-use practices across
groups and time periods were statistically significant. Chi square tests were used to examine differences in
reliance on routine health data between baseline and endline across intervention and control counties, as well
as across key demographic and institutional characteristics. Prior to analysis, the key assumptions for the chi
square test were assessed. Specifically, all variables were categorical, observations were independent and
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2629
www.rsisinternational.org
expected cell counts were examined to ensure that no more than 20 percent of cells had expected frequencies
below five and that no individual cell had an expected count of zero. These requirements were satisfied for all
reported analyses, supporting the validity of the chi square results.
To strengthen causal inference, a Difference in Differences regression model was employed to compare
changes in the outcome variable between intervention and control counties over time. The control group served
as the counterfactual for estimating the net effect of the training. The validity of the DiD approach rests on the
parallel trend’s assumption, which was assessed prior to analysis. These checks showed no systematic
differences in pre intervention slopes or levels between the two groups, providing reasonable support that the
parallel trends assumption was met.
RESULTS
The results present quantitative evidence on how CHMTs apply routine health data within county health
governance structures and how exposure to a structured capacity-building intervention shaped these practices.
The analysis draws on baseline and endline data to provide a coherent account of patterns, variations, and
measurable changes in data-informed decision-making across counties.
A. Socio-Demographic Characteristics of Respondents
Table 1 summarizes the key socio-demographic and professional characteristics of CHMTs members who
participated in the study at baseline and endline across both intervention and control counties. The table
highlights attributes such as gender, age, education level, and duration of CHMT membership, offering insight
into the managerial diversity and institutional experience represented across counties.
Table 1: Socio-Demographic Characteristics of Respondents
Variable
Category
Baseline
Endline
Interven
tion
Control
Interven
tion
Control
Overall
Total Respondents
47.1%
(90)
52.9%
(101)
46.6%
(89)
53.4%
(102)
100.0%
(191)
Age
20 - 29 Years
1.6% (3)
4.7% (9)
1.6% (3)
4.7% (9)
6.3%
(12)
30 - 39 Years
5.8%
(11)
11%
(21)
5.2%
(10)
11%
(21)
16.2%
(31)
40 - 49 Years
25.1%
(48)
21.5%
(41)
25.1%
(48)
21.5%
(41)
46.6%
(89)
50 Years and
Above
14.7%
(28)
15.7%
(30)
14.7%
(28)
16.2%
(31)
30.9%
(59)
Gender
Male
24.6%
(47)
29.8%
(57)
25.1%
(48)
29.8%
(57)
55%
(105)
Female
22.5%
(43)
23%
(44)
21.5%
(41)
23.6%
(45)
45% (86)
Education Level
Diploma
Certificate
4.7% (9)
9.4%
(18)
4.7% (9)
9.4%
(18)
14.1%
(27)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2630
www.rsisinternational.org
Master’s
Degree
13.6%
(26)
15.2%
(29)
13.6%
(26)
15.2%
(29)
28.8%
(55)
PhD
0.5% (1)
2.6% (5)
0.5% (1)
2.6% (5)
3.1% (6)
28.3%
(54)
25.7%
(49)
27.7%
(53)
26.2%
(50)
53.9%
(103)
CHMT Membership
Duration
Less than 1
year
5.8%
(11)
5.8%
(11)
5.8%
(11)
6.8%
(13)
12.6%
(24)
2-5 years
26.7%
(51)
24.1%
(46)
25.7%
(49)
26.2%
(50)
51.8%
(99)
6-9 years
8.4%
(16)
11%
(21)
8.9%
(17)
9.4%
(18)
18.3%
(35)
10 years and
above
6.3%
(12)
12%
(23)
6.3%
(12)
11%
(21)
17.3%
(33)
Males accounted for 54.5% of respondents at baseline and 55% at endline, while females represented 45.5%
and 45%, respectively. The dominant age group was 40–49 years, accounting for about 47% of respondents,
followed by those aged 50 years and above, who comprised roughly 30%. Educational qualifications were
generally high. More than half of all respondents held undergraduate degrees, and about onethird possessed
master’s-level qualifications, underscoring a technically competent and academically prepared workforce.
Regarding tenure, approximately half of the respondents had served between two and five years on their
CHMT, with another one-fifth having over a decade of experience.
B: Descriptive Statistics on Data Use for Common Public Health Decisions
The study also assessed the specific public health decision areas in which County Health Management Teams
reported using routine health data. Table 2 presents the mean agreement scores for each decision domain at
baseline and endline across intervention and control counties. The domains examined include planning and
budgeting, advocacy and resource mobilization, programme monitoring and evaluation, intervention
adjustment, policy and guideline development, and supply chain management.
Table 2: Distribution of Common Public Sector Health Decisions Informed by Routine Health Data
Variable
Baseline
Endline
Interven
tion
Control
Overal
l
Interventio
n
Contro
l
Overall
I use routine health data to design
plans, set priorities, allocate resources,
and develop budgets.
3.72
(0.98)
3.78
(0.88
)
3.75
(0.93
)
4.11
(0.65)
3.59
(0.98
)
3.83
(0.88
)
I rely on routine health data for
advocacy, resource mobilization, grant
applications, and partnerships.
3.76
(0.84)
3.99
(0.77
)
3.88
(0.81
)
4.13
(0.74)
3.97
(0.72
)
4.05
(0.73
)
Routine health data guides my
decisions on continuing, modifying, or
ending interventions and addressing
inequities.
3.84
(0.79)
4.05
(0.74
)
3.95
(0.77
)
4.27
(0.77)
4.01
(0.72
)
4.13
(0.75
)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2631
www.rsisinternational.org
I use routine health data to monitor
and evaluate programme performance,
staff allocation, and training needs.
3.89
(0.89)
4.11
(0.77
)
4.01
(0.84
)
4.35
(0.71)
4.08
(0.79
)
4.20
(0.76
)
Routine health data informs my
decisions on policies, guidelines,
procurement, and distribution of
essential supplies.
3.98
(0.79)
3.90
(0.88)
3.94
(0.84)
4.48
(0.61)
3.85
(0.85)
4.15
(0.81)
Common Decision Made Overall
3.86
(0.73)
3.99
(0.61)
3.93
(0.67)
4.27
(0.42)
3.91
(0.59)
4.08
(0.55)
The results show that routine health data were consistently used to support several core decisions making
functions within county health management. Across both baseline and endline assessments, the most reported
uses of data included planning and budgeting, programme monitoring and evaluation, intervention adjustment,
advocacy and resource mobilization, policy and guideline development, and supply chain management. These
domains recorded mean agreement scores close to or above four at endline in the intervention counties,
indicating strong perceived reliance on data in these areas.
At baseline, overall mean agreement for common decision areas was moderately high in both groups, with
intervention counties reporting 3.86 (SD = 0.73) and control counties 3.99 (SD = 0.61). The most frequently
cited decisions involved planning functions and program monitoring, reflected in the relatively higher baseline
means in these categories. By endline, notable increases were observed across all decision domains in the
intervention counties, where the overall mean rose to 4.27 (SD = 0.42). This was particularly evident in
decisions related to policy development and monitoring and evaluation, which showed the largest gains. In
contrast, the control counties showed minimal change over time, with an endline mean of 3.91 (SD = 0.59).
These patterns suggest that routine data most informed day to day governance tasks such as priority setting,
resource allocation, tracking program performance, adjusting interventions, guiding procurement and
distribution of supplies, and supporting advocacy efforts. The marked improvements in the intervention arm
indicate that capacity strengthening contributed to more systematic and deliberate use of routine data across
these decision domains, while the relative stability in the control arm highlights the absence of comparable
change without structured support.
To complement the quantitative comparisons presented in Table 2, Figures 1 and 2 visualize how reliance on
routine health data shifted over the study period in both intervention and control counties. Figure 1 summarizes
the distribution of CHMT members reporting strong, partial, or limited reliance on routine data for their core
decision making functions, providing a snapshot of how patterns of data use differed between groups at
baseline and endline. Figure 2 presents the corresponding change in mean reliance scores, highlighting overall
shifts in data use intensity across the two study arms. Taken together, the figures illustrate the extent to which
routine data became more consistently integrated into managerial practice following the capacity strengthening
intervention.
Figure 1: Change in Levels of CHMTs Reliance on Routine Health Data for Common Public Sector
Health Decisions
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2632
www.rsisinternational.org
Figure 2: Change in Mean Scores for Reliance on Routine Health Data in By CHMTs for Common
Public Sector Health Decisions
As illustrated in Figures 1 and 2, reliance on routine health data for key public health decisions increased
substantially in the intervention counties while remaining largely unchanged in the control group. At baseline,
75.6 percent of CHMT members in the intervention counties reported strong reliance on routine data compared
to 87.4 percent in the control counties. By endline, this proportion rose to 96.6 percent among the intervention
group, and the limited reliance category was no longer observed. The control counties showed minimal
variation in these categories over the same period.
The pattern is further reflected in the mean reliance scores. In the intervention counties, the mean score
increased from 2.71 at baseline to 2.97 at endline, indicating a measurable strengthening in the consistency
with which routine data were applied to managerial decisions. By contrast, the control counties exhibited
virtually no change, with mean scores shifting only slightly from 2.85 to 2.86. These results highlight a clear
divergence between the two study arms, with intervention counties demonstrating progressively stronger
integration of routine data into decision making processes following the capacity strengthening activities.
C. Inferential Statistics
Inferential analysis using chi-square tests was conducted to determine whether the use of routine health data
for public health decision-making differed significantly across time and between the intervention and control
counties, and whether variations existed across demographic and institutional characteristics. The analysis
addressed the first two study objectives by assessing temporal and cross-sectional differences in data use
patterns and exploring how demographic factors such as gender, education level, tenure, and age intersected
with institutional adoption of evidence-based practices.
Table 3: Chi-square Test for Common Health Sector Decisions Informed by Routine Health Data
Comparison
N
χ2
df
p-value
Control: Baseline vs Endline
203
0.53
2
0.768
Intervention: Baseline vs Endline
179
13.23
2
0.001
Baseline: Control vs Intervention
193
7.85
2
0.02
Endline: Control vs Intervention
189
2.12
2
0.347
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2633
www.rsisinternational.org
Control vs Intervention by gender
382
7.4
4
0.116
Control vs Intervention by Education Level
381
9.62
8
0.292
Control vs Intervention by Duration CHMT
Membership
382
7.2
8
0.515
Control vs Intervention by Age
382
9.13
8
0.331
Results indicated variation in temporal change between the study arms. In the control counties, no statistically
significant difference was observed between baseline and endline, χ² (2, N = 203) = 0.53, p = 0.768, suggesting
that reliance on routine health data remained stable over time in the absence of targeted intervention. In
contrast, the intervention counties recorded a significant shift, χ² (2, N = 179) = 13.23, p = 0.001, reflecting
strengthened integration of data into decision-making processes following capacity-building efforts. Baseline
comparisons showed significant differences between the intervention and control counties, χ² (2, N = 193) =
7.85, p = 0.020, indicating that the two groups did not begin at identical levels of routine data use. This
difference does not in itself threaten internal validity because the Difference in Differences approach explicitly
adjusts for any fixed pre intervention differences between groups. By estimating change over time rather than
relying on baseline levels, the DiD model accounts for uneven starting conditions and isolates the net effect of
the intervention. However, by endline, this variation had dissipated, χ² (2, N = 189) = 2.12, p = 0.347,
indicating that structured training not only enhanced reliance on routine data but also harmonized datause
practices across counties.
Further, the analysis revealed that gender, education, CHMT tenure, and age did not significantly influence
reliance on routine health data, with all p-values exceeding the conventional 0.05 threshold. This pattern
underscores that improvements in the intervention counties were uniformly distributed across demographic and
professional categories rather than confined to specific groups. These findings suggest that structured
interventions can foster a shared professional culture of data-informed governance, where the use of routine
health information becomes an institutional rather than individual attribute.
To strengthen inference, the study also sought to assess the causal effect of the training intervention on the
integration of routine health data into county-level decision-making. A Difference-in-Differences (DiD)
framework was therefore applied, using control counties as the counterfactual to estimate how reliance on data
would have evolved in the absence of the intervention. The DiD model provided a more rigorous estimation of
the intervention’s contribution to institutionalizing data use within health sector governance.
Table 4: DiD Regression Results for Common Health Sector Decisions
Term
Estimate
Std. Error
t-value
p-value
95% CI (Lower,
Upper)
Intercept (Control
Baseline)
3.9821
0.0785
50.70
<0.001
[3.8272, 4.1370]
Intervention (Baseline
Difference)
-0.0712
0.1148
-0.62
0.5360
[-0.2975, 0.1551]
Post (Time Effect in
Control)
-0.0225
0.1110
-0.20
0.8391
[-0.2401, 0.1951]
Intervention × Post
(Treatment Effect)
0.4593
0.1605
2.86
0.0046
[0.1449, 0.7737]
The DiD regression yielded a statistically significant treatment effect of β = 0.4593 (SE = 0.1605, t = 2.86, p =
0.0046, 95% CI [0.1449, 0.7737]). This implies that, on average, exposure to the training intervention
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2634
www.rsisinternational.org
increased the mean reliance score on routine health data by approximately 0.46 units relative to the control
group, after adjusting for time and baseline differences. The intercept (β = 3.9821, p < 0.001) reflected already
strong baseline reliance on data among CHMT members, while the non-significant baseline difference (p =
0.536) confirmed comparability between intervention and control counties prior to the training. The non-
significant time effect in the control group (p = 0.839) indicated that reliance levels remained stable over time
in the absence of structured capacity-building, reinforcing that the observed change was attributable to the
intervention itself.
The findings of the study underscore the transformative role of structured capacity-building in strengthening
the use of routine health data within county-level public health decision-making. Descriptive results
demonstrated clear improvements in data utilization across strategic areas signalling a culture of evidence-
informed management. Inferential analysis further supported this trend, showing that counties exposed to the
intervention demonstrated stronger and more consistent integration of data into managerial processes
compared to those without targeted support. Accordingly, the findings illustrate how deliberate investments in
managerial competencies can elevate data use from an administrative exercise to an institutionalized
governance practice, embedding evidence at the centre of decision-making and fostering greater accountability
and performance within Kenya’s devolved health system.
DISCUSSION
A. Interpretation of Findings
The study demonstrated that structured capacity-building interventions significantly enhanced the integration
of routine health data into public health decision-making at the county level. CHMTs in intervention settings
exhibited a clear shift toward systematic and consistent data use in planning, budgeting, monitoring, and policy
formulation. This pattern signaled a shift away from reporting carried out primarily for compliance purposes
toward a more intentional use of routine data to guide planning, resource allocation, and performance
management. The findings further showed that improvements were broadly distributed across demographic
and institutional characteristics, suggesting that the intervention effectively cultivated a shared culture of data-
driven decision-making. The observed strengthening of data reliance points to the growing recognition of
routine health information not merely as a reporting requirement but as a managerial resource critical to
performance optimization, resource accountability, and adaptive planning within Kenya’s devolved health
framework.
B. Comparison with Existing Literature
These findings align with broader evidence from low- and middle-income countries that emphasize the role of
institutional capacity-building in bridging the gap between data production and use. Studies by Nutley and
Reynolds (2013) and Aqil et al. (2014) similarly highlight that health information systems achieve impact only
when data users are empowered with both technical skills and organizational support. The enhanced data
utilization among CHMTs mirrors outcomes reported in Uganda and Tanzania, where targeted mentorship and
decision-support training substantially improved data-driven planning and supervision (Nsubuga et al., 2018;
Mboera et al., 2020). Moreover, the absence of demographic disparities in data-use practices resonates with
findings by Hotchkiss et al. (2012), who argue that institutional reforms, rather than individual attributes, are
the key determinants of sustained information use.
C. Implications for County Health Governance
The findings have several implications for strengthening Kenya’s devolved health governance, particularly in
promoting systematic and transparent use of routine health data at county level. By showing that structured
capacity building can enhance data use among CHMTs, the study highlights the importance of sustained
investment in health information governance. Counties could reinforce these gains by integrating data use
modules into induction and leadership development programs for senior health managers, ensuring continuity
even during administrative transitions. Moreover, institutionalizing regular data review forums would help
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2635
www.rsisinternational.org
embed evidence informed decision making into routine management processes. For example, counties could
establish monthly CHMT performance review meetings anchored on KHIS dashboards, quarterly
interdepartmental data review sessions focused on priority programs such as RMNCAH or HIV, and annual
county health sector performance dialogues where data guide planning and budgeting decisions. Creating
standardized agendas, templates, and reporting cycles for these forums would further promote consistency and
accountability. The findings suggest that embedding routine data appraisal within existing managerial
structures provides a practical pathway for counties to strengthen collective learning and improve alignment
between evidence and resource allocation.
CONCLUSION AND RECOMMENDATIONS
A. Summary of Conclusions
The study established that structured capacity-building interventions significantly strengthened the integration
of routine health data into county-level decision-making processes in Kenya. CHMTs exposed to targeted
training demonstrated marked improvements in their ability to use data for planning, resource allocation,
programme monitoring, and policy formulation. These findings suggest that strengthening analytical
competencies and providing structured managerial support are associated with increased use of routine data for
planning and performance review, reflecting a shift toward more evidence informed decision making among
CHMTs. Moreover, the uniformity of improvements across gender, educational, and tenure categories
highlights that the intervention’s impact was both inclusive and sustainable. The study concludes that
institutionalizing evidence-based practices through continuous professional development and supportive
governance structures is essential for achieving effective, transparent, and adaptive health system management
within Kenya’s devolved framework.
B. Policy and Practice Recommendations
The findings underscore the need for a national framework to institutionalize the use of routine health data as a
mandatory component of decision-making across all county health departments. The Ministry of Health, in
collaboration with county governments, should embed data-use capacity-building within the continuous
professional development programs for health managers. Counties should also establish regular data review
forums where CHMTs jointly analyse, interpret, and apply evidence to guide planning and budgeting cycles.
Integrating data analytics competencies into leadership training curricula will further ensure that decision-
makers not only access data but can interpret and act on it effectively. Additionally, investment in digital
infrastructure and interoperability between health information systems will enhance the accessibility,
timeliness, and reliability of data required for managerial decision-making. Strengthening accountability
mechanismscan help entrench data use as a governance norm across all administrative tiers.
C. Suggestions for Further Research
Future studies should extend this work by exploring the long-term sustainability of data use practices beyond
the immediate post intervention period. Longitudinal studies could assess whether improvements in data driven
decision making translate into measurable gains in health outcomes, resource efficiency, and service delivery
equity. Further qualitative research is also recommended to examine how organizational culture, leadership
dynamics, and political contexts shape the uptake of data for decision making within devolved systems.
Comparative analyses across sectors such as education, agriculture, or social protection could provide broader
insights into how data use interventions can be scaled and adapted to other governance domains. Equally
important can be sustained research on data use ecosystems will be vital to informing policies that
institutionalize evidence as the foundation for responsive, transparent, and accountable governance across
Kenya’s public service.
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2636
www.rsisinternational.org
ACKNOWLEDGMENT
The author sincerely acknowledges the unwavering guidance and mentorship of Prof. John Paul Oyore from
the Department of Family Medicine, Community Health and Epidemiology, and Prof. George Otieno from the
Department of Health Management and Informatics, both of Kenyatta University. Their intellectual support,
constructive feedback, and commitment throughout the research process were instrumental in shaping the
study’s direction and rigor. The author also extends appreciation to the County Health Management Teams who
participated in the study for their time, insights, and invaluable contribution to advancing evidence-based
decision-making within Kenya’s devolved health system.
REFERENCES
1. AbouZahr, C., &Boerma, T. (2015). Health information systems: The foundations of public health.
Bulletin of the World Health Organization, 83(8), 578–583.
2. Aila, M. A. (2021). Use of routine health information for decision making among health care workers
in Marsabit County, Kenya (Doctoral dissertation, KENYATTA UNIVERSITY).
3. Aqil, A., Lippeveld, T., &Hozumi, D. (2014). PRISM framework: A paradigm shift for designing,
strengthening, and evaluating routine health information systems. Health Policy and Planning, 29(2),
44– 52.
4. Dutta, A., Maina, T., Ginivan, M., & Koseki, S. (2018). Kenya health financing system assessment,
2018: time to Pick the best path. Washington, DC : Palladium, Health Policy Plus.
5. Hotchkiss, D. R., Diana, M. L., &Foreit, K. G. (2012). How can routine health information system data
improve health service delivery? BMC Health Services Research, 12(1), 1–9.
6. Karijo, E. K., Otieno, G. O., & Mogere, S. (2021). Determinants of data use for decision making in
health facilities in Kitui County, Kenya. Quest Journal of Management and Social Sciences, 3(1), 63-
75.
7. Masaviru, L. A., Namusonge, G. S., & Nambuswa, E. M. (2021). Influence of devolved financial
resource governance structure on service delivery in the public health sector in Kenya. Public Policy
and Administration Research, 11(1).
8. Mboera, L. E., Rumisha, S. F., Lyimo, E. P., &Chiduo, M. G. (2020). Data use in decision making for
health services improvement: Lessons from Tanzania. Journal of Health Informatics in Developing
Countries, 14(1), 45–58.
9. Ministry of Health (MoH). (2018). Kenya Health Sector Strategic Plan 2018–2023. Nairobi:
Government of Kenya.
10. Ministry of Health. (2019). National and county health budget analysis FY 2018/19. Ministry of Health.
11. Mutale, W., Chintu, N., Amoroso, C., Awoonor-Williams, K., Phillips, J., Baynes, C., & Sherr, K.
(2013). Improving health information systems for better health outcomes in Zambia. BMC Health
Services Research, 13(1), 58.
12. Muwonge, A., Owuor, C., Williamson, T. S., & Kinuthia, M. (2022). Making devolution work for
service delivery in Kenya. World Bank Publications.
13. Ndabarora, E., Chipps, J., & Uys, L. (2014). Systematic review of health data use in sub-Saharan
Africa. BMC Health Services Research, 14(1), 528.
14. Njuguna, D. C., Muiruri, L., & Njoroge, K. (2022). Utilization of health information data in Nairobi
County public health facilities; lessons from the field. International Journal of Community Medicine
and Public Health, 10(1), 38-44.
15. Nsubuga, P., Kabwama, S. N., &Waiswa, P. (2018). Strengthening district-level decision-making for
health system performance improvement in Uganda. International Journal of Health Policy and
Management, 7(8), 749–758.
16. Nutley, T., & Reynolds, H. (2013). Improving the use of health data for health system strengthening.
Global Health Action, 6(1), 20001.
17. Nutley, T., & Reynolds, H. W. (2013). Improving the use of health data for decision-making: An
intervention framework. Global Health Action, 6(1), 20001.
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XI November 2025
Page 2637
www.rsisinternational.org
18. O’Meara, W. P., Tsofa, B., Molyneux, S., & Goodman, C. (2022). Impact of health systems devolution
on health service delivery in Kenya: A review of the evidence. International Journal of Health Policy
and Management, 11(3), 296–307.
19. Oluoch, T., Santas, X., Kwaro, D., Were, M. C., Biondich, P., Bailey, C., ... &Nzioka, C. (2020). The
impact of routine health information system performance on decision-making in Kenya. BMC Health
Informatics and Decision Making, 20(1), 77.
20. Oware, P. M., Omondi, G., Adipo, C., Adow, M., Wanyama, C., Odallo, D., ... & Were, F. (2025).
Quality and Utilisation of DHIS2 data for health decision making and advocacy in Kenya: A
Qualitative Study. medRxiv, 2025-03.
21. Wako, E., Kimathi, M., & Omondi, B. (2018). Strengthening data demand and use for health system
performance improvement in Kenya. African Health Sciences, 18(4), 1186–1195.
22. World Health Organization (WHO). (2017). Framework and standards for country health information
systems. Geneva: WHO.
23. Zeng, W., Musiega, A., Oyasi, J., Di Giorgio, L., Chuma, J., Lu, R., & Ahn, H. (2022). Understanding
the performance of county health service delivery in Kenya: a mixed-method analysis. Health policy
and planning, 37(2), 189-199.