Cross-Sectional Studies Comprehensive Notes

Cross-Sectional Study Design

Learning Objectives

  • Understand the definition of a cross-sectional study.
  • Understand the purpose of carrying out a cross-sectional study.
  • Learn about sample selection, response rates, measures of outcome and exposure, and the advantages and disadvantages of cross-sectional studies.

Features of a Cross-Sectional Study

  • Takes place at a single point in time.
  • Does not involve manipulating variables.
  • Considers several characteristics at once.
  • Analyzes the prevailing characteristics in a given population.

Definition

  • A cross-sectional study is carried out at one time point or over a short time period, such as a calendar year.
  • It's used to estimate the prevalence of a specified outcome in a defined population.
  • Example: Public health planning.

Data Collection

  • Data can be collected on individual characteristics, including:
    • Exposure to risk factors.
    • Outcome information.
  • Cross-sectional studies provide a picture of the outcome and associated factors at one point in time, offering a 'snapshot'.

Why Conduct a Cross-Sectional Study?

  • When the study purpose is descriptive.
    • Often involves a survey.
    • Often, there is no prior hypothesis.
    • The aim is to describe a population or subgroup with regards to an outcome and a set of risk factors.
  • To determine the prevalence of a specified outcome for a population or subgroups at a particular point in time.
  • Sometimes used to investigate associations between risk factors and the outcome of interest.
  • Limitation: Carried out at only one point in time.
  • Gives no indication of the sequence of events.
  • Cannot determine causality, as it's unclear whether the exposure occurred before, after, or during the disease outcome.
  • However, cross-sectional studies can indicate associations that may exist.
  • Useful for generating hypotheses for further research.
  • Repeated cross-sectional studies can give a pseudo-longitudinal study.
  • Individuals included are chosen from the same or a different sampling frame.
    • Example: British Association for the Study of Community Dentistry Survey.
    • 5-year-old children examined annually.
    • Prevalence of dental caries recorded and monitored over time.
    • Used to inform public health policy and strategies.

Sample Selection & Response Rates

  • A sample frame is used to select the sample.
  • The sample frame and response rate determine how well results can be generalized to the whole population.
  • The sample used is often taken from the whole population.
  • Optimally, the sample is selected using a random method, which is likely to be very representative.
  • For results to represent the population:
    • The selected sample must be representative.
    • Responders need to be representative as well.
  • Nonresponse is a common problem in large surveys.
  • Techniques to minimize nonresponse:
    • Telephone and mail prompting.
    • Second and third mailing of surveys.
    • Letters stressing the importance of replying.
    • Incentives.
  • Biased response is a greater concern.
    • A person is more likely to respond when they have particular characteristic(s).
    • Bias occurs if the characteristic is related to the probability of having the outcome.
    • Example:
      • Door-to-door interview looking for a particular disease.
      • Response rate is highest in the elderly and unemployed, who are more likely to be at home.
      • These groups are also more likely to experience higher levels of the disease, leading to biased results.

Measures of Outcome & Exposure

  • Cross-sectional studies can collect a lot of information about putative risk factors.
  • Loss to follow-up is a common problem in longitudinal studies, where the amount of information collected is often minimized.
  • Loss to follow-up is not an issue in cross-sectional studies.
  • Consider all information that might be relevant.
  • A great opportunity to get a broad knowledge base regarding subjects who have or do not have the specified outcome.
  • It is also important to have optimum response levels.
  • Difficult to confirm associations between outcomes and exposures of long duration using cross-sectional studies.

Advantages

  • Relatively inexpensive.
  • Done in a short time span.
  • Can estimate the prevalence of a specified outcome, as the sample is taken from the whole population.
  • Many outcomes and risk factors can be studied.
  • Used for:
    • Public health planning.
    • Understanding disease etiology.
    • Generation of hypotheses.
  • No loss to follow-up.

Disadvantages

  • Not easy to infer causal associations.
  • Only presents the situation at one point in time, which may change.
  • Prevalence-incidence bias: For long-standing diseases, risk factors associated with death will be under-represented.

Example - Coronary Heart Disease

  • Coronary Heart Disease involves artheroma (plaque) and narrowed artery (atherosclerosis).

Example Study: Coronary Risk Factors

  • Bhopal R et al BMJ 1999;319:215-220
  • Objective: To compare coronary risk factors and disease prevalence among Indians, Pakistanis, and Bangladeshis together with Europeans.
  • Setting: Newcastle upon Tyne.
  • Participants: Men and women, 25-74 years old.
    • 259 Indian
    • 305 Pakistani
    • 129 Bangladeshi
    • 825 European
  • Main outcome measures:
    • Social & economic circumstances.
    • Lifestyle.
    • Self-reported symptoms & diseases.
    • Blood pressure.
    • Electrocardiogram.
    • Anthropometric, hematological & biochemical measurements.
  • Results:
    • Differences in social & economic circumstances, lifestyles, anthropometric measures, and disease.
    • Between Indians, Pakistanis & Bangladeshis and between all South Asians & Europeans.
    • Poorest groups: Bangladeshis & Pakistanis.
  • Most risk factors: Bangladeshis (especially men) fared worst.
    • Smoking most common (57%).
    • Highest concentrations of triglycerides & fasting blood glucose.
    • Lowest concentration of high-density lipoprotein cholesterol.
    • However, blood pressure lowest.
    • Bangladeshis shortest.
  • Higher proportion of Pakistani (22.4%) & Bangladeshi (26.6%) men had diabetes compared to Indians (15.2%).
  • South Asians disadvantaged in a wide range of risk factors.
  • Findings in women were similar to men.
  • Conclusions:
    • Coronary artery risk is not uniform among South Asians.
    • Important differences between Indians, Pakistanis & Bangladeshis for many coronary risk factors.
    • It was previously thought that, apart from insulin resistance, South Asians had lower levels of coronary risk factors than Europeans.
    • This study showed this to be incorrect.
    • May have arisen from combining ethnic subgroups & studying a narrow range of risk factors.

Example – Overweight & Obesity in Children

  • Once a child's BMI is known, it can be plotted on a standard BMI chart.
  • Kids ages 2 to 19 fall into one of four categories:
    • Underweight: BMI below the 5th percentile.
    • Normal weight: BMI at the 5th and less than the 85th percentile.
    • Overweight: BMI at the 85th and below the 95th percentiles.
    • Obese: BMI at or above the 95th percentile.

Example Study: Overweight in Greek Children

  • Hassapidou M et al Hormones (Athens) 2015;14:615-622
  • Objective: Study aimed to assess overweight & obesity prevalence amongst pre-school children in Thessaloniki, Greece.
  • Design: Cross-sectional survey.
    • 1250 pre-school children (657 boys, 593 girls).
    • From state nursery schools in Thessaloniki, Greece.
    • Period 2009-2010.
    • Measurements: Body weight, height, anthropometric, BMI calculated.
  • Results:
    • Rates of excess body weight varied according to international definitions used.
      • Overweight (including obesity): 21.2% - 32.6%.
      • Obesity: 5% - 13.5%.
  • Conclusions:
    • Overweight prevalence is high in pre-school Greek children.

Example – Risk-taking behaviors

Example Study: Risk-Taking in Military Reservists

  • Thandi G et al Occup Med (Lond) 2015;65:413-416
  • Being involved in a military combat role has a negative effect on risk-taking behaviors, e.g., drinking, smoking, risky driving.
  • This study looks at reservists.
  • Aim: To explore the impact of deployment on risk-taking behaviors amongst military reservists.
  • Methods: Cross-sectional study.
    • Self-reported questionnaire to assess hazardous drinking, risky driving, physical violence, smoking, A&E attendance (resulting from risk-taking behaviors).
  • Participants = 1710
  • Response rate = 51%
    • Overall prevalence:
      • Hazardous drinking = 46%.
      • Smoking = 18%.
      • Risky driving = 11%.
      • Attending A&E due to risk behaviors = 13%.
      • Physical violence = 3%.
  • Deployment significantly associated with risky driving, smoking & physical violence.
  • Conclusions:
    • Important to consider the impact of deployment & military factors on the prevalence of risk-taking behaviors.
    • Note that:
      • Study limited: Risk-taking behaviors were examined at a single time point, and the direction of causality cannot be determined.
      • Confounding effects cannot be taken into account: Socio-demographic factors & previous anti-social/violent behaviour.
      • Self-reported behaviors: Social desirability bias.

Example - Insecticidal Net Usage

Example Study: Insecticidal Net Usage in Sri Lanka

  • Whidden CE et al Trans R Soc Trop Med Hyg 2015;109:553-562
  • Long-lasting insecticidal nets distributed in Sri Lanka to control malaria – effectiveness depends on good utilization & maintenance.
  • Aim: To examine the patterns & predictive factors of long-lasting insecticidal nets maintenance and use in Auradhapura district, Sri Lanka.
  • Methods: Cross-sectional study.
    • Data collected & analysed from 530 households.
    • Selected by multi-stage cluster sampling.
    • Data statistically analyzed to identify factors associated with proper maintenance at the household level & with net use the previous night.
  • Results:
    • 377/504 (75%) of households had used all their nets the previous night.
    • 418/504 (82.9%) had used at least one net.
    • 15/474 (3.2%) were maintaining nets to maximize effectiveness.
  • 6 variables associated with use the previous night:
    • More residents.
    • Fewer plain nets.
    • Reporting practical benefits of nets.
    • Conical shape.
    • Newer nets.
    • Lack of side effects.
  • Two variables associated with proper maintenance:
    • Increasing level of education.
    • Taking safety precautions while washing.
  • Conclusions:
    • Practices could improve in places with low malaria transmission if account is taken of recipient preferences.
    • Promotion of long-lasting nets over plain nets.
    • Emphasis on techniques & significance of proper net maintenance.

Example - HPV

  • The 4 most important types of HPV:
    • HPV 16: >83.2% of Cervical Cancer.
    • HPV 18: >50% of Vaginal & Vulvar Cancer
    • HPV 6
    • HPV 11: 90% of Anogenital warts

Example Study: HPV Vaccine Acceptability in China

  • Gu C et al J Clin Nurs 2015;24:2765-2778
  • Aims/Objectives: To examine young women’s perceptions & acceptability of HPV vaccination & factors influencing acceptability in China.
  • Background: HPV vaccines are very important in cervical cancer prevention programmes for young women.
  • Design: Cross-sectional descriptive study in Hunan province, China.
  • Methods: 117 female undergraduates completed a survey in 2012.
    • Five parts to the questionnaire:
      • Background information.
      • Awareness & knowledge of HPV vaccine & cervical cancer.
      • Attitudes to the vaccine & intentions to be vaccinated.
      • Psychosocial burden of HPV infection.
      • HPV-related sexual stigma.
  • Results:
    • Only 44% of participants were willing to be vaccinated.
    • Low awareness & knowledge about the HPV vaccine & cervical cancer.
    • The intention to receive a future vaccination was associated with high levels of knowledge about risk factors for cervical cancer.
  • Conclusions:
    • Low awareness & knowledge among young Chinese women about the value of the HPV vaccine.
    • Social & cultural factors may influence young women’s intentions regarding future vaccination.

Summary

  • Cross-sectional studies are a very useful study design.
  • There are advantages & disadvantages.
  • Relatively cheap.
  • Not time-consuming.
  • Very wide-ranging applicability.
  • However, they can’t be used for causal inference.

Further Reading

  • Levin KA “Study Design III – Cross-sectional studies” Evidence-Based Dentistry 2006; 7:24-25
  • Bhopal R et al “Heterogeneity of coronary heart disease risk factors in Indian, Pakistani, Bangladeshi, and European origin populations: cross sectional study” BMJ 1999;319:215-220
  • Hassapidou M et al “Prevalence of overweight and obesity in preschool children in Tehessaloniki, Greece” Hormones (Athens) 2015;14:615-622
  • Thandi G et al “Risk-taking behaviours among UK military reservists” Occup Med (Lond) 2015;65:413-416
  • Whidden CE et al “Patterns and predictive factors of long-lasting insecticidal net usage in a previously high malaria endemic area in Sri Lanka: a cross-sectional survey” Trans R Soc Trop Med Hyg 2015;109:553-562
  • Gu C et al “Human papillomavirus vaccine acceptability among female undergraduate students in China: the role of knowledge and psychosocial factors” J Clin Nurs 2015;24:2765-2778
  • Webb P, Bain C, Page A. Essential Epidemiology: An Introduction for Students and Health Professionals. Cambridge University Press; Third Edition