Design and validation of a questionnaire to assess household-level socioeconomic status: a mixed-methods study

Background

The systematic assessment of socioeconomic status (SES) of individual households in relation to health outcomes is considered the most important step to tackle inequalities in health systems. There is no valid and reliable questionnaire to evaluate household-level SES in low-to-middle-income countries. This study aimed to design and validate a questionnaire to assess household-level SES.

Methods

The study used a three-stage sequential mixed methods design. First, we did a systematic review to identify and select a comprehensive set of SES indicators and variables from conceptual frameworks and models. PubMed, Scopus, Emerald, Elsevier, Ovid, Google Scholar, Springer, ProQuest, WHO, and Word Bank were searched for studies published in English between Jan 1, 1990, and March 1, 2015. Second, the initial questions were written in such a way as to be precise and unambiguous. Third, we assessed content validity with the content validity ratio (CVR) and content validity index (CVI). Additionally, a qualitative approach was applied to assess face validity for ambiguity, relevancy, and difficulty. Finally, test–retest reliability was used to test intraclass correlation coefficient.

Findings

From an initial yield of 248 papers, 21 studies met our inclusion criteria. We extracted 30 SES indicators, and finally selected 20 indicators during the internal discussion and through experts' opinions by use of the nominal group technique. The questionnaire was designed in eight sections with 79 questions (household identifying information, n=10; household supervisor demographic and general information, 9; household location, housing, and living facilities, 6; household food habits, 7; household expenditure and income, 20; physical activity, 3; health services use, 12; social capital, 12). Face validity was confirmed by six experts. CVI for all questions was acceptable (≥0·85), CVR for 76 questions was acceptable (0·71–1·00), and three questions were modified. Test–retest correlation coefficients were acceptable for all questions (0·43–0·89).

Interpretation

We have demonstrated that a newly developed questionnaire to assess household-level SES in low-to-middle-income countries meets various criteria of validity and reliability. The questionnaire will allow us to gather more comprehensive information on household-level SES in different countries and settings to make informed decisions about the impact of policies and reforms on the SES.

Funding

None.