Methods: We conducted a scoping review to map prevalence surveys conducted in LMICs published between 1995 and 2018. We followed Arksey and O'Malley's six-step framework. The search was conducted in OVID Medline, EMBASE, ISI Web of Science, Global Health, WHO Global Index Medicus and included three domains: CRDs, prevalence and LMICs. After an initial title sift, eight trained reviewers undertook duplicate study selection and data extraction. We charted: country and populations, random sampling strategies, CRD definitions/phenotypes, survey procedure (questionnaires, spirometry, tests), outcomes and assessment of individual, societal and health service burden of disease.
Results: Of 36 872 citations, 281 articles were included: 132 from Asia (41 from China). Study designs were cross-sectional surveys (n = 260), cohort studies (n = 11) and secondary data analysis (n = 10). The number of respondents in these studies ranged from 50 to 512 891. Asthma was studied in 144 studies, chronic obstructive pulmonary disease (COPD) in 112. Most studies (100/144) based identification of asthma on symptom-based questionnaires. In contrast, COPD diagnosis was typically based on spirometry findings (94/112); 65 used fixed-ratio thresholds, 29 reported fixed-ratio and lower-limit-of-normal values. Only five articles used the term 'phenotype'. Most studies used questionnaires derived from validated surveys, most commonly the European Community Respiratory Health Survey (n = 47). The burden/impact of CRD was reported in 33 articles (most commonly activity limitation).
Conclusion: Surveys remain the most practical approach for estimating prevalence of CRD but there is a need to identify the most predictive questions for diagnosing asthma and to standardise diagnostic criteria.
Methods: We administered relevant translations of the BOLD-1 questionnaire with additional questions from ECRHS-II, performed spirometry and arranged specialist clinical review for a sub-group to confirm the diagnosis. Using random sampling, we piloted a community-based survey at five sites in four LMICs and noted any practical barriers to conducting the survey. Three clinicians independently used information from questionnaires, spirometry and specialist reviews, and reached consensus on a clinical diagnosis. We used lasso regression to identify variables that predicted the clinical diagnoses and attempted to develop an algorithm for detecting asthma and COPD.
Results: Of 508 participants, 55.9% reported one or more chronic respiratory symptoms. The prevalence of asthma was 16.3%; COPD 4.5%; and 'other chronic respiratory disease' 3.0%. Based on consensus categorisation (n = 483 complete records), "Wheezing in last 12 months" and "Waking up with a feeling of tightness" were the strongest predictors for asthma. For COPD, age and spirometry results were the strongest predictors. Practical challenges included logistics (participant recruitment; researcher safety); misinterpretation of questions due to local dialects; and assuring quality spirometry in the field.
Conclusion: Detecting asthma in population surveys relies on symptoms and history. In contrast, spirometry and age were the best predictors of COPD. Logistical, language and spirometry-related challenges need to be addressed.