Mapping Pathways to Early Lung Cancer Diagnosis in South Africa: A Mixed-Methods Protocol for Community-Engaged Intervention Development

Patient-level barriers to timely diagnosis

Limited public awareness of lung cancer symptoms constitutes a primary driver of diagnostic delay.^7,11 Studies across SSA demonstrate that patients frequently fail to associate persistent cough or unexplained weight loss with malignancy, leading to prolonged symptom appraisal periods.^7,8,12 Fear of cancer diagnosis, stigma surrounding smoking-related disease, and reliance on traditional healers further discourage early healthcare-seeking behaviour.¹¹ Socioeconomic constraints compound these barriers. Many South Africans—particularly in rural areas—face prohibitive travel costs to access healthcare facilities.^8,11 Gender dynamics also shape healthcare-seeking patterns: women may prioritise family responsibilities over medical attention, whilst men often delay care due to stoicism and peer pressure that frames healthcare-seeking as weakness.¹³

Health system barriers and diagnostic pathway fragmentation

South Africa’s healthcare system is characterised by stark public-private disparities. The public sector serves 80% of the population but remains critically under-resourced, with insufficient specialist capacity and protracted diagnostic waiting times.¹⁴ In contrast, the private sector—accessible only to the 27% with medical insurance—offers advanced diagnostic technologies beyond the reach of most citizens.^15,16 Within the public sector, primary care clinics conducting most initial consultations often lack diagnostic tools and trained personnel to assess suspected lung cancer cases promptly. Many primary care providers have limited exposure to lung cancer and may misattribute symptoms to more prevalent conditions such as tuberculosis or chronic obstructive pulmonary disease.^7,8 Knowledge gaps regarding lung cancer symptomatology, risk factors, and referral protocols—documented in KwaZulu-Natal studies—result in misdiagnosis, unnecessary antibiotic prescriptions, and delayed specialist referral.^7,8

Even when lung cancer is suspected, diagnostic pathways remain fragmented.⁷ Access to imaging and histopathology services varies substantially, with some regional hospitals lacking basic radiology capabilities.¹¹ Biopsy turnaround times are prolonged by workforce shortages and logistical constraints in transporting specimens to centralised laboratories.¹¹ Each delay permits further disease progression, reducing the probability of curative interventions such as surgery or radiotherapy.⁷

Environmental context: air pollution and lung cancer risk

Air pollution represents an underestimated contributor to South Africa’s lung cancer burden. Global models frequently underestimate South African air quality challenges, creating misleading perceptions.¹⁷ Urban and industrial areas exhibit elevated particulate matter (PM2.5) concentrations—particularly in densely populated low-income communities—linked directly to increased lung cancer risk.¹⁸ The Highveld region routinely exceeds health-based ambient air quality standards.¹⁷ Approximately 70% of low-income households depend on solid fuels (wood, coal) for cooking and heating, whilst dust storms, unpaved roads, agricultural activities, vehicle emissions, and coal-fired power stations further degrade air quality.¹⁸ This environmental burden intersects with healthcare system weaknesses: climate-related extreme weather events—exemplified by the 2022 cyclones affecting 2.8 million Southern Africans—disrupt already fragile health infrastructure, creating additional barriers to cancer diagnosis and treatment.^18–22

Study rationale and aim

Addressing lung cancer diagnostic delays in South Africa requires moving beyond documenting barriers towards developing implementable, contextually appropriate solutions. Despite clear evidence of multi-level obstacles—patient awareness deficits, provider knowledge gaps, system fragmentation, and environmental risk factors—no comprehensive mapping of diagnostic pathways exists for South African settings. Furthermore, the MRC framework for complex interventions emphasises that effective intervention development requires robust evidence synthesis and theory development before proceeding to co-design and pilot testing.

This mixed-methods protocol provides that foundational evidence base through systematic pathway mapping. Work Package 1 maps lung cancer diagnostic pathways in the Western Cape by: (i) synthesising international LMIC evidence through realist review; (ii) quantifying time intervals and delay patterns via cross-sectional survey; (iii) exploring lived experiences of patients, families, and healthcare workers through journey mapping; and (iv) assessing diagnostic capacity across facility levels. Participatory feedback workshops will engage stakeholders in developing multi-level action plans, identifying priority intervention targets and establishing the empirical foundation for subsequent co-design work (Work Package 2) aligned with MRC complex interventions guidance.

Study setting and healthcare context

This study will be conducted in the Western Cape Province, South Africa. The South African public healthcare system operates across three tiers: primary care clinics staffed predominantly by nurses, secondary facilities (district/regional hospitals) providing diagnostic investigations, and tertiary teaching hospitals offering specialist services. Patients with suspected cancer typically self-present to primary care, receive initial assessment, and are referred upwards depending on clinical findings and facility capacity. The public sector serves approximately 80% of the population but faces substantial resource constraints—insufficient specialist workforce, prolonged waiting times, and variable diagnostic infrastructure.^14,15,23 By contrast, the private sector (accessible to 27% with medical insurance) consumes half of national health expenditure yet remains financially prohibitive for most citizens.^15,16

Theoretical and methodological framework

This protocol integrates three complementary frameworks. The Model of Pathways to Treatment²⁴ structures our empirical data collection, defining key time intervals (patient interval, primary care interval, pre-diagnostic interval, total diagnostic interval) and multi-level factors influencing diagnostic journeys. The Medical Research Council framework for complex interventions^25,26 positions this work as the evidence synthesis and theory development phase, generating findings to inform subsequent co-design work. Participatory action research principles^27,28 underpin stakeholder engagement throughout, ensuring community voices shape problem definition, data interpretation, and solution generation. Reporting follows SPIRIT-PRO guidance for protocols, GRAMMS for mixed-methods studies, COREQ for qualitative components, and STROBE for cross-sectional surveys.

Study design

This 24-month mixed-methods study comprises four integrated components designed to triangulate quantitative, qualitative, and system-level data on lung cancer diagnostic pathways in the Western Cape.

Component 1: Realist Review of Lung Cancer Diagnosis in LMICs

Rationale: To synthesise international evidence on diagnostic strategies in resource-constrained settings, identifying context-mechanism-outcome configurations (CMOCs) applicable to South Africa.

Methods: We will conduct a realist review following established methodologies,^29,30 focusing on how, why, for whom, and in what circumstances diagnostic interventions succeed or fail in LMICs. The review examines generative causality—the underlying mechanisms connecting interventions to outcomes within specific contexts. Initial programme theories will be iteratively refined through stakeholder engagement and emerging evidence. The full protocol has been published separately.³¹

Integration with WP1: CMOCs identified through realist synthesis will inform survey instrument development (Component 2), provide comparative contexts for journey mapping interpretation (Component 3), and establish international benchmarks for facility capacity assessment (Component 4).

Component 2: Cross-Sectional Survey of Diagnostic Time Intervals

Objective: To quantify patient, primary care, pre-diagnostic, and total diagnostic intervals; identify sociodemographic and clinical factors associated with diagnostic delay; and establish baseline data for future intervention evaluation.

Study Design: Hospital-based cross-sectional survey conducted at two tertiary-level teaching hospitals and appropriate secondary-level facilities in the Western Cape capable of confirming lung cancer diagnosis.

Participants

Inclusion criteria: Adults aged ≥18 years with either (i) histologically or cytologically confirmed lung cancer diagnosed within the preceding four weeks, or (ii) high clinical suspicion of lung cancer defined as radiological findings (chest X-ray or CT demonstrating pulmonary nodules/masses ≥8 mm) plus clinical features (haemoptysis, unexplained weight loss >5%, persistent cough >3 weeks unresponsive to antibiotics) warranting multidisciplinary team discussion or tissue diagnosis.

Exclusion criteria: Individuals unable to provide informed consent due to cognitive impairment, severe illness precluding interview, or language barriers not resolvable through available interpreters.

Sampling and Sample Size

Consecutive patients who provide written informed consent will be recruited over 12 months at each site. In the absence of facility-level incidence data, we adopt a pragmatic approach: assuming 40–60 new lung cancer diagnoses annually per tertiary site (based on Western Cape Cancer Registry estimates), a 50% consent rate, and participation from two tertiary and three secondary facilities, we anticipate recruiting 100–150 participants. This will provide median time interval estimates with acceptable precision (95% CI ±7–10 days for median patient interval of 30 days, assuming interquartile range of 14–60 days). If recruitment falls below projections, the data collection period will be extended to 18 months.

Participants

Patients and families: Up to 50 newly diagnosed lung cancer patients (within four weeks of diagnosis or treatment plan development) or patients with high clinical suspicion of lung cancer, recruited from participating facilities. When patients attend with family members/caregivers, dyadic interviews will be conducted after written informed patient consent. Purposive maximum-variation sampling will ensure representation across age, sex, urban/rural residence, disease stage, and facility level. Recruitment will continue until thematic saturation is achieved across key demographic strata. Thematic saturation will be assessed iteratively after every 10 interviews using constant comparison method. Saturation is defined as no new codes emerging across three consecutive interview sets. If saturation occurs before n = 50, recruitment will cease; if not achieved by n = 50, recruitment will extend to n = 60 with interim analysis justification documented.

Healthcare workers: 12–15 clinicians and managers involved in lung cancer diagnosis across primary, secondary, and tertiary levels. Purposive sampling will target: primary care nurses (n = 3), general practitioners (n = 2), respiratory physicians (n = 3), thoracic surgeons (n = 2), radiologists (n = 2), clinical managers (n = 3). Participants will be recruited following completion of patient interviews to allow healthcare worker mapping sessions to respond to emerging patient journey findings.

Methodological considerations

Three methodological challenges warrant acknowledgement. First, facility-based recruitment cannot capture patients who never reach diagnostic services—the ‘denominator problem’ documented across sub-Saharan Africa where cancer registry coverage ranges from 2.3% to 100%.⁴² Second, patient recall of symptom timelines introduces measurement error; whilst triangulation with medical records improves accuracy,⁴³ record completeness in under-resourced settings constrains this strategy. Third, whilst co-designed interventions show improved acceptability, evidence of superior clinical outcomes remains limited.⁴⁴ We accept these constraints whilst emphasising WP1’s descriptive mapping aim rather than causal inference or definitive intervention testing.

Strengths and limitations

Key strengths include comprehensive multi-level data collection integrating patient, provider, and system perspectives; explicit theoretical framing facilitating replication; and participatory feedback workshops operationalising knowledge translation. The realist review component situates South African findings within international LMIC evidence, illuminating context-dependent versus transferable diagnostic strategies.

Limitations warrant acknowledgement. Generalisability beyond the Western Cape is constrained by inter-provincial variation. Gilson et al. (2024) identified the Western Cape as a “pocket of relative bureaucratic effectiveness” with sustained service delivery reform.⁴⁵ Significant variations exist even within the province between metropolitan and remote areas.⁴⁶ Our findings may not apply to less-resourced provinces with weaker infrastructure. Exclusive focus on public-sector facilities excludes the 20% accessing private care, where diagnostic pathways differ substantially. Cross-sectional survey design precludes causal inference; associations between sociodemographic factors and diagnostic intervals may be confounded by unmeasured variables (disease biology, tumour location, histological subtype). Framework analysis prioritises pragmatic actionability over phenomenological depth, trading epistemological richness for translational relevance.

Policy implications and dissemination

Findings will directly inform three policy audiences. The Western Cape Department of Health will receive facility-specific capacity assessments identifying resource gaps. South Africa’s developing national cancer plan will access diagnostic pathway data supporting evidence-based recommendations for clinician education and referral protocol standardisation. Primary care training institutions will integrate lung cancer symptom recognition into curricula, addressing provider knowledge gaps.

Evidence on knowledge translation effectiveness in LMIC cancer research remains limited. Edwards et al. (2019)⁴⁷ found policy briefs, deliberative dialogues, and capacity-building workshops most commonly used, but most studies report knowledge outcomes rather than clinical impact.^47,48 Only 21 studies across 20 LMICs examined interventions targeting diagnostic timeliness, with many lacking clinically relevant measures.⁴⁹ We address this gap through multi-channel dissemination: peer-reviewed open-access publications; briefs for health management teams; infographics for community health workers; webinars for primary care networks; and feedback sessions at participating facilities.

Internationally, the protocol contributes methodological precedent for LMIC cancer diagnostic research. Low-resource settings often lack infrastructure for large-scale cohort studies or randomised trials. Our pragmatic data collection within existing clinical services demonstrates how mixed-methods approaches can generate policy-relevant evidence. The integration strategy offers a replicable template for other cancer types or non-communicable diseases requiring multi-step diagnostic pathways.

Conclusion

Reducing lung cancer diagnostic delays in South Africa demands precise understanding of where, why, and for whom delays occur. This protocol generates that understanding through rigorous mixed-methods inquiry, positioning our intervention development on an empirical foundation. By integrating patient voice, provider perspectives, and system-level capacity assessment within participatory research principles, we create conditions for contextually appropriate, implementable, and sustainable solutions. The study advances lung cancer control in South Africa whilst offering methodological insights transferable to other LMIC cancer and non-communicable disease diagnostic challenges.