Household Medicine Use: Methodologies and Analysis

Discover a comprehensive review offering critical insights into the design and execution of household surveys on medicine utilization, essential for researchers, policymakers, and public health professionals.

Understanding how communities access, afford, and utilize medicines from the consumer’s perspective is paramount for developing effective health policies and evaluating health systems. However, collecting accurate data on medicine use, especially in low and middle-income countries where computerized systems are scarce, presents significant methodological challenges. Self-reported data, gathered through direct interviews, is prone to reporting bias, making a carefully designed methodology crucial.

Bertoldi et al.’s seminal 2008 article, “A descriptive review of the methodologies used in household surveys on medicine utilization,” published in BMC Health Services Research, addresses these challenges head-on. This systematic review provides a vital resource by describing the methodological and analytical aspects of quantitative household studies on medicine utilization, analyzing employed methodologies qualitatively, and offering actionable recommendations for future research in the field. Its findings are indispensable for researchers aiming to improve the quality of medicine use data and to navigate the complexities of study planning and analysis.

Key Highlights of the Review:

The authors conducted a rigorous systematic review of original papers published between 1995 and 2008, where the household was a sampling unit. Searching Medline/Pubmed, Scielo, and Lilacs, and meticulously applying specific inclusion and exclusion criteria, they identified 61 relevant studies.

The review reveals several key trends and methodological practices among medicine utilization studies:

• Geographic Focus and Design: The majority of studies (64.0%) were conducted in Europe and North America, predominantly employing a cross-sectional approach.
• Data Collection Methods: Over 80% of studies relied on face-to-face interviews for data collection.
• Recall Period: The most frequently used recall period was 14–15 days, though the range varied broadly from one day to two years [3, 16, Figure 1]. Studies on chronic diseases often asked about current use without a defined reference period, while others differentiated recall periods for acute versus chronic medicines.
• Data Validation: A significant proportion (59%) of studies trained interviewers to request medicine packaging from subjects, enhancing data accuracy. However, medical prescriptions were requested far less frequently (15%).
• Questionnaire Design: Approximately 75% of studies utilized open-ended questions to investigate medicine use. The Anatomical Therapeutic Chemical (ATC) classification system was the most common for pharmacological groups, though 42.5% of studies did not specify their classification system.
• Analytical Denominators: Most studies used individuals as the denominator in their analyses.

Crucial Methodological Insights and Recommendations for Future Studies:

The authors delve into critical methodological aspects, providing invaluable guidance:

• Optimizing Recall Period: While there’s no single “ideal” recall period, the authors highlight its profound impact on reported prevalence and accuracy. They suggest that 14–15 days is a recommended period for comparison with existing literature. Longer periods can lead to underreporting, particularly for self-prescribed medicines, as recall accuracy decreases.
• Choosing Administration Mode: Face-to-face interviews are strongly recommended, especially in low and middle-income countries, due to their higher response rates and the practical limitations (e.g., low telephone coverage, high illiteracy) of other methods like telephone or postal surveys.
• Crafting Effective Questions: Researchers should use simple, technically correct terminology and clearly define “medicine use” to respondents. Special attention should be paid to prompting for seemingly “less important” medicines like analgesics, natural products, or topical treatments, which are often underreported. Open-ended questions are preferred for evaluating overall medicine utilization.
• Utilizing Proxy Information: For studies involving elderly subjects, who may have higher medicine utilization but difficulty providing accurate information, interviewing close relatives or caregivers (proxies) is recommended to avoid underestimating prevalence.
• Leveraging Packaging and Prescriptions: Requesting to see medicine packaging is a highly valuable strategy for validating data and aiding recall in household studies. While less common, requesting prescriptions can further qualify and validate collected data.
• Standardizing Classification and Denominators: Authors are urged to explicitly state the pharmacological group classification system used (e.g., ATC) and the denominator (individuals vs. medicines) employed in analyses, as these choices significantly affect data interpretation and comparability across studies.

Conclusion:

This comprehensive review serves as a critical update for researchers on current methodological standards in household medicine utilization surveys. It not only summarizes existing practices but also provides thoughtful recommendations that can significantly improve the quality of data obtained and guide the planning and analysis of future studies. The authors advocate for greater standardization of methods and more detailed reporting to facilitate international comparisons and replication.

A closer look at the review

Bertoldi and colleagues searched PubMed, SciELO and LILACS for household-based studies of medicine use published between January 1995 and June 2008, in Portuguese, English or Spanish. Studies in hospitals, schools or workplaces, studies of a single drug group and studies limited to one disease were excluded. Of 4,852 records, 61 studies met the criteria. Two authors judged eligibility and coded each study for design, age group, sample size, sampling strategy, mode of administration, respondent type, recall period, question format, whether packaging or prescriptions were checked, drug classification system and denominator.

What the 61 surveys looked like

Most surveys were cross-sectional and 64% were carried out in Europe or North America. More than 80% used interviewers and 95.1% used random sampling. The most common recall period was 14 to 15 days, but periods ranged from one day to two years. About three quarters asked open-ended questions, 59% asked to see medicine packages and 15% asked for prescriptions. Reporting was weak: 93.4% gave no information on the validity of the questionnaire and 42.5% did not name the classification system used. Among surveys of older people, only 26.3% used proxy respondents.

The choice of recall period matters. In two surveys from Pelotas, Brazil, antimicrobial use among adults was 7.3% with a 30-day recall and 5.4% with a 15-day recall, a difference large enough to change policy conclusions.

Practical recommendations

The authors recommend a recall period of around 14 days to keep results comparable, a clear and simple definition of “medicine” so that analgesics, natural products and topical preparations are captured, and face-to-face interviews, especially in low- and middle-income countries. Open-ended questions suit measures of overall use, while checklists suit specific drugs or conditions. Checking packages and using proxy informants improves data for older adults. Every survey should report its classification system, denominator and exact questions.

Why it matters for health services research

Household surveys are often the only way to see medicine use outside the health system, including self-medication. The review shows that methodological differences, rather than real differences in behaviour, can explain much of the variation between studies, and that standardised methods and fuller reporting are needed before survey results can guide pharmaceutical policy.

Reference for the Article:

Bertoldi, A. D., Barros, A. J. D., Wagner, A., Ross-Degnan, D., & Hallal, P. C. (2008). A descriptive review of the methodologies used in household surveys on medicine utilization. BMC Health Services Research, 8(1), 222. https://doi.org/10.1186/1472-6963-8-222