Explaining the attitude-behaviour gap in adolescent sexual and reproductive health service utilization: A cognitive dissonance perspective from rural Uganda

Abstract

Adolescents in low-resource settings hold positive attitudes toward sexual and reproductive health (SRH) services, yet fail to use them. This study presents a theoretical reinterpretation of mixed-methods data (N = 408 adolescents, aged 13–19) from Uganda’s Busoga region, through Festinger’s Cognitive Dissonance Theory. Quantitative findings showed health-seeking attitudes predicted SRH service utilization (β = 0.910, p < .001) and future intentions (β = 1.976, p < .001), but 52% of adolescents had no SRH service contact in past year, and mean uptake scores remained below 2.0 on a 5-point scale. Qualitative analysis of focus group discussions identified three dissonance-reduction strategies: avoidance of health facilities, rationalization through misinformation, and selective information-seeking from peers. The reinterpretation proposes four intervention strategies: barrier reduction, community norm campaigns, cues to action with immediate access, and targeted messaging for working and rural adolescents. Addressing psychological discomfort and structural barriers may help close the attitude-behaviour gap.

Introduction

The adolescent sexual and reproductive health (SRH) paradox

Uganda’s adolescent population (aged 10 to 19 years) constitutes approximately 24.8% of the national population (Uganda Bureau of Statistics, 2024). This demographic group faces elevated risks of unintended pregnancy, sexually transmitted infections (STIs), and HIV acquisition, yet SRH service utilization remains persistently low, particularly in rural eastern regions such as Busoga (Tumwizere et al., 2025; Uganda Bureau of Statistics and ICF, 2023). Studies consistently document a paradox: adolescents report favourable attitudes toward SRH services, endorsing their importance, confidentiality, and accessibility, but rarely seek care (Kakal et al., 2022; Nabugoomu et al., 2020).

This inconsistency between favourable attitudes and limited utilization represents an attitude-behaviour gap that is not unique to adolescent SRH. It has been documented across diverse health domains, including vaccination (Brewer et al., 2017), contraceptive adherence (Frost et al., 2012), and HIV testing (Strauss et al., 2015). However, the magnitude of the gap in adolescent SRH has significant public health consequences: delayed care-seeking contributes to preventable morbidity, adolescent pregnancy, and HIV transmission (Patton et al., 2016; Sully et al., 2020). Although previous research has extensively documented barriers to adolescent SRH service utilization, comparatively little attention has been devoted to understanding the psychological processes through which adolescents reconcile positive attitudes with persistent non-use.

Theoretical framing: Beyond traditional health behaviour models

The Health Belief Model (HBM) (Rosenstock, 1974) conceptualizes health actions as functions of perceived susceptibility, severity, benefits, and barriers. Andersen’s Behavioural Model of Health Services Use (Andersen, 1995) emphasises predisposing, enabling, and need factors. Previous analyses of the data examined in this study employed both frameworks and found that attitudes, perceptions, and preferences significantly predicted utilization (Nyende et al., 2025). However, these frameworks may not fully account for the psychological and motivational processes that contribute to the persistence of the attitude–behaviour gap.

Festinger’s Cognitive Dissonance Theory (CDT) (Festinger, 1957) provides a complementary explanatory perspective. Festinger proposed that individuals experience psychological discomfort, which he termed dissonance, when they hold two or more contradictory cognitions (beliefs, attitudes, or behaviours). The magnitude of dissonance is a function of the importance of the cognitions, the strength of the inconsistency, and the degree of external justification for the discrepant behaviour. To reduce discomfort, individuals may change their behaviour, alter their attitudes, or add consonant cognitions that justify the inconsistency (Festinger, 1957; Harmon-Jones and Mills, 2019).

In the present study, the core dissonant pair is proposed as a conceptual application of CDT rather than an established construct within the literature. Specifically, it refers to the coexistence of two psychologically inconsistent cognitions:

• • Cognition A (Positive Attitude): “SRH services are valuable, confidential, and should be used by someone like me.”

• • Cognition B (Behavioural Reality): “I have not accessed any SRH service in the past 12 months.”

According to CDT, maintaining these two cognitions simultaneously may generate psychological discomfort that motivates adolescents to reduce the inconsistency. Importantly, dissonance does not necessarily result in behaviour change. Individuals frequently restore psychological consistency by adopting alternative cognitive or behavioural strategies that allow existing behaviour to continue (Festinger, 1957; Harmon-Jones and Mills, 2019). Resolution strategies observed in health behaviour research include avoidance of dissonance-arousing information, rationalization of non-use, and selective exposure to consonant information (Cooper, 2019; Kahan, 2017; McGrath, 2017). These mechanisms are among the best-established processes described in CDT and have been documented across numerous health behaviours, including smoking cessation, vaccination, HIV prevention, and condom use (Freijy and Kothe, 2013; Jolley and Douglas, 2017; Levy et al., 2021). The directed qualitative analysis undertaken in this study was therefore organized around these theoretically established dissonance-reduction mechanisms.

Although CDT offers a potentially useful framework for interpreting the observed attitude-behaviour gap, alternative explanations should also be considered. Structural vulnerability and bounded agency perspectives (Harmon-Jones and Mills, 2019; Mendelsohn et al., 2014) suggest that highly structured environments and economic constraints can limit the expression of individual agency, shaping adolescents’ decisions and aspirations within the barriers and contexts in which they find themselves. Stigma theory provides another lens: Hall et al. (2018) demonstrated that SRH stigma, enacted through gossip, marginalisation, and mistreatment from healthcare workers and community members, directly precludes adolescents’ use of family planning methods and services. Social norms theory similarly suggests that collective norms around contraceptive use at the community level are significantly associated with individual contraception use among adolescents (Sedlander and Rimal, 2019). Recent studies in Uganda have confirmed that peer norms and community expectations strongly influence adolescents’ SRH decision-making, often overriding individual attitudes and intentions (Kakal et al., 2022; Nabugoomu et al., 2020). Additionally, the attitude-behaviour gap itself has been extensively studied in health psychology, with meta-analyses demonstrating that experimentally induced changes in attitudes lead to medium-sized changes in intention and behaviour. Yet, the gap between intention and action remains a persistent challenge (Sheeran et al., 2016). In low-resource settings, behaviour may be constrained less by internal inconsistency than by economic, social, and institutional limitations (Jewkes et al., 2014). Recent research in sub-Saharan Africa (SSA) has expanded the understanding of the barriers to adolescent SRH service utilization. A systematic review by Chandra-Mouli et al. (2019) identified provider attitudes, lack of privacy and confidentiality, cost, and distance as persistent barriers across the region. In the Ugandan context, recent studies have documented multiple dimensions of the attitude–behaviour gap. Tumwizere et al. (2025) found that among sexually active adolescent girls and young women in Busoga, non-utilization of modern contraceptives was associated with fear of side effects, distance to facilities, and partner disapproval. A mixed-methods study by Akatukwasa et al. (2019) in southwestern Uganda revealed that adolescents frequently cited fear of being seen at health facilities and concerns about confidentiality as reasons for avoiding SRH services. Ninsiima et al. (2021) documented how gender norms and power dynamics within relationships constrain adolescent girls’ ability to negotiate condom use and seek SRH services in rural Uganda. These recent findings underscore the persistence of the attitude–behaviour gap and highlight the need for theoretical frameworks that can explain the psychological processes underlying non-utilization despite positive attitudes.

The present study

This study re-examines existing mixed-methods data through a theoretical lens. The original study was not designed to test dissonance theory; rather, the current analysis retrospectively applies dissonance concepts to explain the observed gap between attitude and behaviour. The specific objectives are:

• 1. To quantify the attitude-behaviour gap across multiple SRH service types using descriptive statistics from the original survey

• 2. To identify qualitatively expressed dissonance-reduction strategies in focus group discussion transcripts

• 3. To propose theory-consistent intervention strategies for converting positive attitudes into SRH service utilization

Theoretical framework: Cognitive dissonance theory

Festinger (1957) defined cognitive dissonance as an aversive motivational state arising from the perception that one’s beliefs, attitudes, or behaviours are inconsistent with each other. The theory has been extensively developed and applied across psychological domains (Harmon-Jones and Mills, 2019).

As introduced earlier, the core dissonant pair in adolescent SRH juxtaposes positive attitudes toward SRH services against the behavioural reality of non-use. The specific resolution strategies examined in this study are avoidance of dissonance-arousing situations, rationalization through the addition of consonant cognitions, and selective information-seeking from peers. These three strategies were selected a priori, ahead of data analysis, because they correspond to the principal dissonance-reduction mechanisms identified in Festinger (1957) original formulation and because they have each received independent empirical support in subsequent health behaviour research (Kahan, 2017); other mechanisms described in the broader dissonance literature, such as direct attitude change or behaviour change, were not examined here because the cross-sectional design could not capture change over time.

Dissonance in health behaviour

Cognitive dissonance has been productively applied to a range of health behaviours. In smoking cessation, dissonance-based interventions have been shown to reduce post-cessation regret and maintain abstinence (Freijy and Kothe, 2013). In vaccination research, parents who wrote counter-attitudinal essays about vaccine safety subsequently reported higher intentions to vaccinate (Jolley and Douglas, 2017). More recently, Pearce and Cooper (2021) demonstrated that a cognitive dissonance intervention (advocating for safety measures while recalling personal inconsistencies) increased compliance with COVID-19 guidelines, including greater intentions to seek vaccination, highlighting the applicability of dissonance principles to contemporary health challenges.

In sexual health, the classic demonstration of dissonance-based intervention comes from Stone et al. (1994), who showed that young adults who publicly advocated for safe sex and were then reminded of their own past failures subsequently purchased more condoms, a pattern explained through cognitive dissonance processes.

More recent evidence confirms that cognitive dissonance continues to play a role in condom use interventions. Levy et al. (2021) conducted a randomised controlled trial examining psychological inoculation as a strategy to increase condom use tendencies. They found that participants in the psychological inoculation condition showed higher cognitive dissonance scores compared to control participants, and the intervention was effective at increasing condom use tendencies among those who had sexual relations. The authors explicitly noted that their psychological inoculation intervention “targeted a specific underlying process (i.e., cognitive dissonance)” (Levy et al., 2021: 588).

Despite these successful applications, cognitive dissonance theory has not been systematically applied to adolescent SRH service utilization in SSA. Most dissonance-based health interventions have focused on individual behaviours such as condom use, smoking, or vaccination, rather than on the decision to seek services from health facilities. The present study addresses this gap by applying dissonance theory retrospectively to existing mixed-methods data from the Busoga region, examining whether the same dissonance-reduction strategies observed in other health domains (avoidance, rationalization, and selective information-seeking) operate in the context of adolescent SRH service utilization.

Methods

Data source

This is a secondary analysis of data from a doctoral mixed-methods study conducted in Bugweri and Iganga districts, Busoga region. Study ethics approval was obtained and registered with the relevant bodies. Written informed consent was secured from all adult participants and from parents/guardians of minors, with adolescents’ assent. Participation was voluntary, and confidentiality and anonymity were maintained.

The original study employed a convergent, concurrent mixed-methods design, in which the quantitative cross-sectional survey and the qualitative focus group discussions (FGDs) were collected during the same field period from overlapping communities, then analyzed separately before being merged for interpretation. The full sampling frame, instrument development, psychometric validation, and statistical diagnostics for the quantitative strand are reported in the original study (Nyende et al., 2025), to which the reader is referred for methodological detail beyond the summary provided here.

Quantitative sample and measures

Multistage stratified cluster sampling yielded 408 adolescents (51.5% female, 58.1% rural) aged 13–19 years, with stratification ensuring representation by sex and rural–urban residence. Measures included health-seeking attitudes (11 items, α = 0.85) assessed using a five-point Likert scale, and SRH service uptake (12 types, α = 0.86). Additional covariates included age, sex, residence, educational status, working status, travel time to the nearest health facility, and access to health information. Analyses used Stata 18 (descriptive statistics, ordered logistic regression, and structural equation modelling). Model fit for the structural equation models was evaluated using the Root Mean Square Error of Approximation (RMSEA), the Comparative Fit Index (CFI), the Tucker-Lewis Index (TLI), and the Standardized Root Mean Square Residual (SRMR) (Browne and Cudeck, 1992; Hu and Bentler, 1999). For this secondary analysis, the original statistical outputs were reinterpreted through the lens of CDT rather than re-estimated. Model covariates and full diagnostics are available from the corresponding author.

Qualitative sample and re-analysis

The original study included four FGDs with 38 adolescents, stratified by sex and residence, to capture diverse perspectives (Krueger and Casey, 2015). Table 1 summarises the composition.

Table 1.

Composition of focus group discussions.

FGD	Gender	Residence	Age range	Number of participants
1	Female	Urban (Bugweri)	13 to 17	9
2	Female	Rural (Iganga)	13 to 17	9
3	Male	Urban (Iganga)	17 to 19	10
4	Male	Rural (Bugweri)	17 to 19	10

Each FGD was facilitated by a trained moderator using a semi-structured discussion guide exploring adolescents’ attitudes toward SRH services, barriers to service use, sources of SRH information, and experiences with health facilities. The discussions lasted 60 to 90 minutes and were conducted in Lusoga. Discussions were audio-recorded, transcribed verbatim, and translated into English.

For the current re-interpretation, transcripts were re-analyzed using directed content analysis (Hsieh and Shannon, 2005) with a coding framework informed by CDT (Festinger, 1957). Directed content analysis is particularly appropriate where an existing theoretical framework guides interpretation of qualitative data while allowing new insights to emerge inductively.

The analysis did not seek to directly measure cognitive dissonance; rather, it explored whether participants’ narratives could be meaningfully interpreted through concepts derived from dissonance theory. Three a priori coding categories were developed from Festinger’s original theory and subsequent developments by Harmon-Jones and Mills (2019):

• • avoidance of dissonance-arousing situations;

• • rationalization through the addition of consonant cognitions; and

• • selective information-seeking that reinforces existing beliefs.

These categories were selected as shown in Table 2.

Table 2.

Coding framework for dissonance-reduction strategies.

Dissonance-reduction strategy	Operational definition
Avoidance	Deliberately avoiding dissonance-arousing situations or information
Rationalization	Adding consonant cognitions that justify non-use
Selective information-seeking	Seeking information that justifies non-use and avoiding information that would increase dissonance

Two independent coders coded the transcripts; discrepancies were resolved through discussion. Inter-coder agreement was 87% (Cohen’s kappa = 0.82) (Cohen, 1960). To enhance trustworthiness, findings were triangulated across data sources (quantitative and qualitative) and participant perspectives (sex, residence, age) (Nowell et al., 2017). Importantly, excerpts were interpreted as illustrating behaviours consistent with CDT rather than constituting direct evidence that participants consciously experienced cognitive dissonance.

Mixed-methods integration approach

For this secondary analysis, a convergent integration approach was employed (Creswell and Plano Clark, 2017). Quantitative data from the survey and qualitative data from focus group discussions were analysed separately and then integrated at the interpretation stage to provide a comprehensive understanding of the attitude–behaviour gap. The quantitative findings provided population-level evidence of the gap’s magnitude and the factors predicting utilization, while the qualitative findings provided insights into the psychological processes underlying non-use. Integration occurred through: (1) joint display of quantitative and qualitative findings in the results section; (2) narrative weaving, where qualitative data were used to illustrate and explain quantitative patterns; and (3) identification of convergent and complementary evidence across data sources (Fetters et al., 2013).

Results

This section first summarises the core theoretical constructs that frame the findings (Table 3), then presents quantitative evidence of the attitude–behaviour gap and its predictors, and finally presents qualitative evidence of the dissonance-reduction strategies adolescents used to manage that gap. Table 3 provides an overview of the principal concepts from CDT that informed the interpretation of both the quantitative and qualitative findings. Rather than presenting empirical results, it serves as an analytical framework for understanding how inconsistencies between adolescents’ favourable attitudes towards SRH services and their utilization behaviours may generate psychological discomfort and motivate dissonance-reduction strategies.

Table 3.

Core constructs of cognitive dissonance theory applied to adolescent SRH.

Construct	Definition	Application to adolescent SRH
Dissonance magnitude	The degree of discomfort, increasing with the importance of the cognitions and the strength of the inconsistency	Greater dissonance when adolescents strongly endorse SRH importance, yet have never used services
External justification	Sufficient external reasons for behaviour that reduce perceived personal responsibility	Weak external justification when barriers (stigma, distance) are perceived as modifiable rather than inevitable
Dissonance reduction	Strategies to restore consistency: Behaviour change, attitude change, or addition of consonant cognitions	Observed strategies in this study include avoidance, rationalization, and selective information-seeking

Note. Definitions adapted from Festinger (1957) and Harmon-Jones and Mills (2019).

Quantitative evidence of the attitude-behaviour gap

Adolescents demonstrated favourable behavioural readiness for SRH engagement. Table 4 presents the mean scores for each readiness construct.

Table 4.

Behavioural readiness indicators among adolescents.

Construct	Mean	SD	Interpretation
Health-seeking attitudes	3.65	0.65	Moderate–high readiness
Perceptions of SRH services	3.38	0.67	Mixed but positive
Preferences for facility care	4.04	0.54	High willingness
Composite behavioural readiness	3.69	—	Overall moderate–high readiness

Note. The composite score represents the mean of the three psychosocial readiness constructs. All scores are on a 1 to 5 scale.

As shown in Table 4, adolescents reported moderate-to-high health-seeking attitudes (mean = 3.65) and high preferences for facility care (mean = 4.04). Perceptions of SRH services were somewhat lower (mean = 3.38) but remained above the scale midpoint. The composite behavioural readiness score of 3.69 suggests that adolescents were psychologically prepared to engage with SRH services. Despite these positive attitudes, service uptake was markedly low.

Table 5 categorizes adolescents by intensity of SRH service use over the past year.

Table 5.

Distribution of SRH service use (N = 408).

Utilization category	N (%)	Interpretation
No SRH use in the past year	214 (52.5%)	Indicates potential unmet need
Occasional use	173 (42.4%)	Transitional engagement
Regular use	21 (5.1%)	Established utilization behaviour

Over half of adolescents (52.5%) reported no SRH service use in the past year, indicating substantial unmet need.

Table 6 presents the attitude-behaviour gap across SRH service types, showing both the percentage who never used each service and the corresponding mean uptake scores.

Table 6.

Attitude-behaviour gap across SRH service types (N = 408).

Service type	Percent “Never” used (12 months)	Mean uptake score (1–5)	Interpretation
Contraceptive services	70.0%	1.56	Very low
HIV testing services	66.4%	1.52	Very low
HIV prevention services	79.7%	1.42	Very low
Youth-friendly clinic services	66.2%	1.58	Very low
Mobile health services	78.4%	1.34	Very low
Online SRH resources	78.4%	1.38	Very low
Pharmacy contraceptive supplies	73.0%	1.43	Very low
SRH counselling services	54.2%	1.82	Low to moderate
Sexual education sessions	48.4%	1.90	Low to moderate
Radio/TV SRH resources	33.9%	2.22	Moderate
Peer SRH education/support	40.4%	2.22	Moderate
School/community SRH services	43.1%	2.04	Moderate
Composite mean (all 12 services)	–	1.68	Low

Note. Mean scores are on a 5-point scale (1 = Never, 5 = Very Often).

Table 6 reveals a substantial gap between adolescents’ positive attitudes and their actual service use. Across all 12 service types, the composite mean uptake score was only 1.68, well below the scale midpoint of 3.0. The lowest uptake was observed for mobile health services (mean = 1.34), online SRH resources (mean = 1.38), and HIV prevention services (mean = 1.42). Even the most commonly used services such as radio/TV SRH resources (mean = 2.22) and peer SRH education/support (mean = 2.22) remained below the midpoint.

Health-seeking attitudes as predictors of SRH utilization

Ordered logistic regression confirmed that health-seeking attitudes significantly predicted both past SRH visit frequency and future utilization intentions. Tables 7 and 8 present the ordered logistic regression results for past SRH visit frequency and future utilization intentions, respectively.

Table 7.

Ordered logistic regression of health-seeking attitudes on frequency of SRH visits (past 12 months).

Variable	Coefficient	SE	Z	p	95% CI
Health-seeking attitudes	0.910	0.17	5.23	<.001	[0.57, 1.25]

Note. McFadden’s pseudo R² = 0.034. N = 404.

Table 8.

Ordered logistic regression of health-seeking attitudes on future SRH service utilization.

Variable	Coefficient	SE	Z	p	95% CI
Health-seeking attitudes	1.976	0.18	10.94	<.001	[1.62, 2.33]

Note. McFadden’s pseudo R² = 0.142. N = 404.

As shown in Tables 7 and 8, health-seeking attitudes significantly predicted both past utilization (coefficient = 0.910, p < .001) and future utilization intentions (coefficient = 1.976, p < .001). Notably, attitudes were stronger predictors of future intentions than of past behaviour, suggesting that while adolescents intend to use services, they face barriers to translating these intentions into action. The modest pseudo R² values (0.034 for past behaviour, 0.142 for future intentions) indicate that attitudes alone explain only a small proportion of the variance in actual behaviour.

Structural barriers as dissonance triggers

Adolescents reported several structural barriers that may impede access to SRH services. Table 9 summarises the proportion of participants reporting each barrier.

Table 9.

Proportion of adolescents reporting key structural barriers to SRH service use (N = 408).

Barrier	N (%)	Dissonance function
Distance to facility (≥30 min travel time)	221 (54.2%)	Provides external justification for non-use
Fear of judgement from providers/community	244 (59.8%)	Justifies avoidance of dissonance-arousing situations
Lack of privacy/confidentiality in services	76 (18.6%)	Rationalizes non-use as self-protection
Cost or time burden of seeking care	189 (46.3%)	Adds consonant cognition

Note. Multiple responses were permitted; percentages are based on the total sample size.

Table 9 shows that fear of judgement from providers or community (59.8%) and distance to facilities (54.2%) were the most commonly reported barriers. Consistent with CDT, such barriers may function not only as practical obstacles but also as external justifications that reduce the psychological discomfort of non-use.

Direct effects of psychosocial factors and barriers on SRH utilization

Structural equation modelling was used to examine the direct relationships between psychosocial factors, structural characteristics, and SRH service utilization outcomes. Overall model fit was acceptable (RMSEA = 0.082; CFI = 0.923; TLI = 0.839; SRMR = 0.023), indicating that the proposed model adequately represented the observed data.

Table 10 presents the direct effects from the structural equation model. These coefficients show which factors directly influence SRH utilization.

Table 10.

Direct effects from the structural equation model.

Predictor	Outcome	Coefficient	Std. Error	p-value	95% CI
Health-seeking perceptions	Current utilization	−0.239	0.082	0.004	[−0.399, −0.078]
Travel time to facility	Current utilization	−0.271	0.083	0.001	[−0.448, −0.094]
Health-seeking attitudes	Future utilization	2.310	0.453	<0.001	[1.422, 3.198]
Age	Future utilization	0.348	0.106	0.001	[0.141, 0.555]
Working status	Future utilization	−1.822	0.626	0.004	[−3.049, −0.596]

Note. *p < 0.01, **p < 0.001. Coefficients are from structural equation models.

Table 10 reveals several important direct relationships. Negative perceptions of SRH services (coefficient = −0.239, p = 0.004) and longer travel time to facilities (coefficient = −0.271, p = 0.001) significantly reduced current utilization.

Conversely, positive attitudes (coefficient = 2.310, p < 0.001) significantly increased future utilization intentions but did not translate into current action. Age (coefficient = 0.348, p = 0.001) also positively predicted future intentions, with older adolescents expressing stronger intentions to use services. However, working status (coefficient = −1.822, p = 0.004) negatively predicted future intentions. This finding is explored further in the Discussion in relation to structural barriers as external justifications for non-use.

Moderated effects of structural conditions on SRH utilization

Table 11 presents the moderated (interaction) effects.

Table 11.

Moderated effects from the structural equation model.

Interaction	Outcome	Coefficient	Std. Error	p-value	95% CI
Perceptions × Age	Current utilization	0.208	0.037	<0.001	[0.136, 0.280]
Preferences × Travel time	Current utilization	0.940	0.225	<0.001	[0.612, 1.269]
Attitudes × Radio access	Current utilization	0.072	0.146	<0.001	[0.043, 0.101]
Attitudes × Working status	Future utilization	−0.452	0.147	0.002	[−0.741, −0.165]
Preferences × Working status	Future utilization	0.857	0.156	<0.001	[0.552, 1.162]
Preferences × Travel time	Future utilization	−0.042	0.009	<0.001	[−0.059, −0.024]
Attitudes × age	Future utilization	−0.091	0.024	0.002	[−0.147, −0.034]

Note. Interaction coefficients from moderated SEM. Positive coefficients indicate amplification of the main effect; negative coefficients indicate attenuation.

*p < 0.01, **p < 0.001.

Table 11 reveals several significant interactions between psychosocial factors and structural conditions. The positive interaction between perceptions and age (coefficient = 0.208, p < 0.001) indicates that as adolescents grow older, their perceptions of SRH services become a stronger predictor of actual use.

The positive interaction between preferences and travel time (coefficient = 0.940, p < 0.001) shows that adolescents with strong service preferences overcame the distance barrier. The negative interaction between preferences and travel time for future intentions (coefficient = −0.042, p < 0.001) suggests that while preferences help overcome distance for current use, the burden of travel may still dampen future intentions, particularly for adolescents who have experienced the costs of travelling to facilities.

The positive interaction between attitudes and radio access (coefficient = 0.072, p < 0.001) indicates that exposure to health information via radio helped adolescents translate positive attitudes into actual service use.

The negative interaction between attitudes and working status (coefficient = −0.452, p = 0.002) shows that employment weakens the relationship between positive attitudes and future utilization intentions. However, the positive interaction between preferences and working status (coefficient = 0.857, p < 0.001) indicates that employed adolescents with strong service preferences still intend to use services.

Qualitative evidence of dissonance-reduction strategies

Directed content analysis of focus group discussion transcripts identified three dissonance-reduction strategies, consistent with Festinger’s original formulation, that adolescents employed to manage the attitude–behaviour gap. Each strategy corresponds to one of the three a priori coding categories. This section integrates illustrative participant quotes to demonstrate how adolescents articulate these strategies in their own words.

Strategy 1: Avoidance

Participants frequently described avoiding SRH services because they anticipated judgement, embarrassment, breaches of confidentiality, or other negative interactions with health workers. These concerns were reported by both male and female adolescents across rural and urban settings. One female participant stated:

“The health workers are sometimes tough, and that makes us fear opening up and telling them everything that is really happening in our lives.” (Female adolescent, FGD 2, rural Iganga)

A male participant described how fear of disclosure led to avoidance:

“Some of us are shy to ask, and that stops us from going.” (Male adolescent, FGD 4, rural Bugweri)

Another female participant articulated the anticipated consequences of seeking care:

“If you are pregnant and scared to tell your parents, you decide to go and tell the health worker, but she starts blaming you and asking why you engaged in things you cannot manage.” (Female adolescent, FGD 2, rural Iganga)

These accounts highlight how anticipated negative experiences led adolescents to avoid health facilities, thereby preventing the situation in which the inconsistency between positive attitudes and non-use would become salient.

Strategy 2: Rationalization

Participants added consonant cognitions that justified non-use of SRH services. Common rationalizations involved misinformation and negative beliefs about contraceptives and condoms. One female participant stated:

“Sometimes I hear that family planning methods may cause you to become barren, and that scares me.” (Female adolescent, FGD 2, rural Iganga)

A male participant expressed a similar belief:

“Use of family planning by a woman who has never had children makes her sterile.” (Male adolescent, FGD 3, urban Iganga)

Adolescents also expressed fear about condom safety. One female participant stated:

“Even condoms sometimes remain inside you and can lead to serious problems or even death.” (Female adolescent, FGD 2, rural Iganga)

Another male participant elaborated:

“About condoms… we are told they affect women. They can get stuck in women, and that the lubricant in condoms is dangerous.” (Male adolescent, FGD 3, urban Iganga)

Adolescents also expressed concern that SRH services were not designed for them:

“They don't give adolescents enough attention, especially on condoms.” (Female adolescent, FGD 2, rural Iganga)

These beliefs, though medically inaccurate, appeared to help adolescents reconcile positive general attitudes toward SRH services with their personal non-use.

Strategy 3: Selective information-seeking

Adolescents expressed a strong preference for obtaining SRH information from peers rather than from health workers. This selective approach allowed them to access information that was more likely to support their existing beliefs and behaviours. One female participant stated:

“During conferences, health workers should bring in youth to educate fellow youth.” (Female adolescent, FGD 1, rural Bugweri)

The same participant elaborated on why peer information is valued:

“Although we listen to health workers, we tend to believe more in our peers when they tell us something.” (Female adolescent, FGD 1, rural Bugweri)

Another participant described the use of digital media for SRH information as follows:

“We use phones to search for information when we need it.” (Male adolescent, FGD 3, urban Iganga)

Coexistence of contradictory cognitions

Several participants expressed both positive views of health workers and clear barriers to accessing services within the same discussion. For example, one male participant stated:

“I feel health providers give the right information about SRH, so it's good to talk to them.” (Male adolescent, FGD 4, rural Bugweri)

Another male participant expressed trust in health workers:

“When I need advice on sexual issues, I go to a health worker, and I trust that health workers give the best advice regarding sexual issues.” (Male adolescent, FGD 4, rural Bugweri)

Yet the same participants faced access barriers. One noted:

“Sometimes we need condoms, but we cannot access them, not even at health facilities, and that makes it hard to protect yourself when you need to.” (Male adolescent, FGD 4, rural Bugweri)

Together, these accounts illustrate the cognitive inconsistency that characterizes the attitude–behaviour gap: participants voiced strong trust in health workers while simultaneously describing barriers that prevented them from acting on that trust.

Discussion

This study argues that the persistent gap between adolescents’ favourable attitudes toward SRH services and their low utilization is not simply a product of structural obstacles, but is actively maintained by psychological processes that recruit those obstacles as justifications for non-use. By applying CDT to mixed-methods data from the Busoga region, Uganda, the study demonstrates that avoidance, rationalization, and selective information-seeking operate alongside structural barriers to sustain the attitude-behaviour gap. This motivational dimension has been largely absent from existing explanatory frameworks for adolescent SRH service utilization in low-resource settings, and its recognition represents the central theoretical contribution of this study.

Principal findings

The findings reveal a pronounced attitude-behaviour gap: adolescents generally expressed favourable attitudes, perceptions, and preferences towards SRH services, yet utilization remained consistently low. More than half of participants reported no contact with SRH services during the previous year despite positive psychosocial readiness. While favourable attitudes significantly predicted both current utilization and future intentions, the relatively modest explanatory power of the regression models suggests that positive attitudes alone are insufficient to explain adolescents’ health-seeking behaviour.

The qualitative findings help explain this discrepancy. Adolescents described avoiding health facilities because of anticipated judgement and confidentiality concerns, endorsing misconceptions regarding contraception and condoms, and relying primarily on peers rather than health professionals for SRH information. When interpreted alongside the quantitative evidence showing positive attitudes but low utilization, these accounts suggest that psychological processes may operate alongside structural barriers to maintain the observed attitude-behaviour gap. Although the present study did not directly measure cognitive dissonance, these qualitative patterns are consistent with avoidance, rationalization, and selective information-seeking, which are recognised mechanisms through which individuals may reduce cognitive inconsistency (Festinger, 1957; Harmon-Jones and Mills, 2019).

Structural barriers as external justifications for non-use

A central finding of this study is that structural barriers serve not only as objective obstacles to SRH service access but also as powerful external justifications that reduce cognitive dissonance and perpetuate non-use. This dual function of barriers has not been systematically examined in previous adolescent SRH research, which has typically treated barriers as purely practical obstacles (Chandra-Mouli et al., 2019).

This interpretation is consistent with Festinger (1957) assertion that individuals with strong external reasons for their behaviour experience less dissonance because they can attribute their actions to factors beyond their control. The moderated effects in Table 11 provide empirical support for this proposition: barriers that provide strong external justifications (distance, employment) attenuate the relationship between positive attitudes and service use, while factors that increase personal responsibility (age, strong preferences, media access) amplify it.

The practical implication is that interventions focusing solely on information provision or attitude change may be insufficient if structural barriers remain in place. When barriers are removed, adolescents lose their external justifications for non-use, increasing the psychological discomfort of the attitude–behaviour gap and motivating behaviour change. This suggests that barrier reduction should be prioritized as a dissonance-reduction strategy in its own right, rather than merely as a practical improvement to service access.

Interplay between stigma, social norms, and dissonance

The qualitative findings reveal how stigma and social norms interact with dissonance processes. Adolescents’ anticipation of judgement from health workers and community members functions as both a practical barrier and a dissonance-reduction mechanism. The fear of being labelled as promiscuous or irresponsible, as captured in the quote, “she starts blaming you and asking why you engaged in things you cannot manage”, provides a powerful external justification for avoidance (Hall et al., 2018; Mbachu et al., 2021).

Similarly, social norms around peer influence shape selective information-seeking. Adolescents’ preference for peer information over health worker information reflects both a practical preference for accessible sources and a psychological strategy to maintain cognitive consistency (Kahan, 2017; McGrath, 2017). Peers are more likely to validate non-use and reinforce rationalizations, while health workers are more likely to challenge misinformation and recommend service use.

The interplay between stigma, norms, and dissonance suggests that multi-level interventions are needed.

Extending CDT to adolescent SRH service utilization in low-resource settings

These findings extend previous applications of cognitive dissonance to health behaviour in three important ways. First, dissonance-based interventions have been successfully applied to individual health behaviours such as smoking cessation (Freijy and Kothe, 2013), vaccination uptake (Jolley and Douglas, 2017), condom use (Stone et al., 1994), and more recently, COVID-19 safety behaviours (Pearce and Cooper, 2021). Yet, despite this breadth, this study represents one of the first systematic applications of CDT to adolescent SRH service utilization in a low-resource setting. The identification of the core dissonant pair (positive attitudes toward SRH services versus non-use) provides a theoretical foundation for understanding why adolescents who endorse SRH services fail to use them. Second, the identification of three distinct resolution strategies (avoidance, rationalization, and selective information-seeking) provides a more comprehensive account than prior studies that have treated dissonance reduction monolithically.

The avoidance strategy aligns with findings on health-seeking avoidance in stigmatised conditions (Strauss et al., 2015).

Third, the findings demonstrate that CDT complements, rather than replaces, traditional frameworks such as the HBM and Andersen’s Behavioural Model. Where these models explain utilization as a function of perceived barriers and enabling resources, CDT explains the motivational dynamics that sustain non-use even when barriers are recognised. The two perspectives are synergistic: structural barriers create the conditions for dissonance, and dissonance-reduction strategies perpetuate non-use.

Intervention implications

Persistent gaps between adolescents’ knowledge and positive attitudes toward SRH services and their actual utilization are well documented in the broader literature. Studies conducted in eastern Uganda have shown that adolescents frequently recognise the benefits of contraception, HIV testing, and counselling while simultaneously reporting fear of judgement, misinformation, restrictive social norms, and limited service accessibility (Kakal et al., 2022; Nabugoomu et al., 2020; Tumwizere et al., 2025). Similar findings have been reported elsewhere in SSA, where structural barriers and perceived stigma continue to discourage adolescents from seeking reproductive healthcare despite increasing awareness of available services (Jacobs et al., 2023; Okyere et al., 2024). The present study extends this literature by proposing a theoretical mechanism through which these barriers may interact with adolescents’ psychological decision-making. Building on existing evidence on dissonance-based interventions and structural interventions for adolescent SRH, four complementary strategies are proposed.

Strategy 1: Reduce structural barriers to enable consonant behaviour

Removing key structural barriers, including distance, cost, provider judgement, and lack of confidentiality, can help adolescents translate favourable attitudes into actual service use. Evidence from Uganda indicates that mobile health services and school-linked programmes effectively reduce distance barriers (Kakal et al., 2022), while training health workers in adolescent-friendly communication has been shown to reduce fear of judgement (Chandra-Mouli et al., 2019). Improving confidentiality through dedicated adolescent hours or private consultation spaces is particularly important, given that 18.6% of participants in this study reported lack of privacy as a barrier.

Strategy 2: Add accurate consonant cognitions through community norm campaigns

Community norm campaigns and peer-led messaging can normalize SRH service utilization and directly counter common misinformation about contraception and condoms. Interventions such as the Get Up Speak Out! programme in Uganda have demonstrated the effectiveness of peer-led approaches in shifting social norms and increasing SRH knowledge (Kakal et al., 2022). The specific misconceptions identified in this study, that family planning causes infertility and that condoms are dangerous, should be explicitly and repeatedly addressed through multiple channels, including radio, peer networks, and school-based programmes.

Strategy 3: Amplify cues to action with immediate low-cost access

Dissonance-based interventions have proven effective in other health domains through hypocrisy paradigms, in which individuals publicly advocate for a behaviour and are then reminded of their own inconsistencies (Stone et al., 1994; Strauss et al., 2015). In the adolescent SRH context, such campaigns could encourage public endorsement of service use, followed immediately by opportunities to act through mobile clinics or youth-friendly facilities. More recently, Pearce and Cooper (2021) demonstrated that asking individuals to advocate for protective health behaviours while reflecting on their own inconsistencies significantly increased intentions to comply with COVID-19 safety guidelines, confirming that the hypocrisy paradigm remains effective across diverse health contexts.

Strategy 4: Target subgroups with the largest attitude–behaviour gaps

Interventions for working adolescents could include workplace-based SRH services, flexible opening hours, and mobile units that visit employment sites (Chandra-Mouli et al., 2019). For rural adolescents, strategies should focus on mobile outreach, reducing travel time, and addressing misinformation that tends to be more prevalent in remote areas (Ninsiima et al., 2021; Tumwizere et al., 2025).

Strengths and limitations

A major strength of this study is the integration of quantitative and qualitative data within a coherent theoretical framework, allowing behavioural patterns to be examined from multiple perspectives. The mixed-methods design strengthened the credibility of the findings through methodological triangulation, while independent coding of the qualitative data enhanced analytical rigour. The study was conducted in a high-need setting (Busoga region) with a relatively large quantitative sample (N = 408), increasing its relevance for policy and programming in similar low-resource contexts.

Several limitations should be acknowledged. First, this represents a secondary theoretical reinterpretation of data that were not originally collected to test CDT. To mitigate this, the analysis explicitly acknowledges that the original study was not designed to measure cognitive dissonance and frames the findings as a theoretically informed interpretation rather than a confirmatory test. The consistency between quantitative and qualitative findings, together with the use of both deductive (theory-derived) and inductive (data-driven) coding strategies, further enhances the credibility of the interpretation (Lincoln et al., 1985).

Second, the cross-sectional design limits causal inference regarding the temporal relationships between attitudes, dissonance processes, and service utilization. To partially address this, the study drew on both past and future utilization measures, allowing for a more nuanced understanding of potential temporal patterns.

Third, the study was conducted in two districts within the Busoga region and may not be fully generalisable to other parts of Uganda or SSA. To enhance transferability, the findings are situated within the broader literature on adolescent SRH across the region, noting both consistencies and contextual differences.

Finally, social desirability bias remains a possibility in self-reported data. This was mitigated through the use of mixed methods, anonymous data collection procedures, and focus group discussions conducted in small, same-sex groups, which may have reduced individual pressure to provide socially desirable responses.

Conclusion

Adolescents in the Busoga region of Uganda hold favourable attitudes toward SRH services, yet utilization remains low. Applying CDT as a complementary lens, this study suggests that the attitude–behaviour gap is maintained through three dissonance-reduction strategies: avoidance of services, rationalization through misinformation and negative beliefs, and selective information-seeking from peers. Structural barriers appear to function not only as practical obstacles but also as external justifications that reduce psychological discomfort.

Strategies that reduce barriers, normalize service use and counter misinformation via community norm campaigns, provide immediate access following cues to action, and target high-gap subgroups (working and rural adolescents) warrant further development and testing.

Future research should prospectively measure cognitive dissonance and evaluate CDT-informed interventions in randomised designs.

ORCID iD

Noah Robert Nyende https://orcid.org/0009-0001-6342-4113

Ethical considerations

Ethical approval for this study was obtained from the Mildmay Uganda Research Ethics Committee (MUREC-2024-451), followed by registration with the Uganda National Council for Science and Technology (SS3374ES). Administrative permission was also secured from the district authorities of Bugweri and Iganga.

Consent to participate

Written informed consent was obtained from all participants (with parental consent and adolescent assent for minors). Further details are provided in the Methods section.

Consent for publication

Written informed consent for publication of anonymised quotations and de-identified data was obtained from all participants (with parental consent for minors).

Data Availability Statement

De-identified data generated and/analyzed during this study are available from the corresponding author on request.*