From text to codings: intercoder reliability assessment in qualitative content analysis.

BACKGROUND: High intercoder reliability (ICR) is required in qualitative content analysis for assuring quality when more than one coder is involved in data analysis. The literature is short of standardized procedures for ICR procedures in qualitative content analysis. OBJECTIVE: To illustrate how ICR assessment can be used to improve codings in qualitative content analysis. METHODS: Key steps of the procedure are presented, drawing on data from a qualitative study on patients' perspectives on low back pain. RESULTS: First, a coding scheme was developed using a comprehensive inductive and deductive approach. Second, 10 transcripts were coded independently by two researchers, and ICR was calculated. A resulting kappa value of .67 can be regarded as satisfactory to solid. Moreover, varying agreement rates helped to identify problems in the coding scheme. Low agreement rates, for instance, indicated that respective codes were defined too broadly and would need clarification. In a third step, the results of the analysis were used to improve the coding scheme, leading to consistent and high-quality results. DISCUSSION: The quantitative approach of ICR assessment is a viable instrument for quality assurance in qualitative content analysis. Kappa values and close inspection of agreement rates help to estimate and increase quality of codings. This approach facilitates good practice in coding and enhances credibility of analysis, especially when large samples are interviewed, different coders are involved, and quantitative results are presented.

Clarifying gender interactions in multivariate analysis.

OBJECTIVES: To examine a linear regression model to predict physical fitness using an alternative concept to analyse gender interactions. METHODS: Data were obtained from the Berne Lifestyle Panel, a survey on health and lifestyles of 56-66 years old Bernese citizens. A measure of physical fitness was regressed on gender, education and their interaction as central explanatory variables and age as confounding factor. For ease of interpretation, two dummy variables of education are introduced, one for female, the other for male education. The model with education dummy variables is compared to a linear regression model without interaction stratified by gender, and with a model with the multiplicative gender-education interaction term without stratification. RESULTS: The use of dummy variables ensures an accurate description of both women's and men's associations of education with the dependant variable, without losing any explanatory power due to stratification. CONCLUSION: The results show that the use of dummy variables is a rewarding alternative to stratification and conventional gender interaction analysis, providing both sufficient statistical information and a basis for straightforward interpretation.

Physician gender and changes in drug prescribing after the implementation of reference pricing in British Columbia.

BACKGROUND: Gender-specific attitudes and communication styles are known to influence both the content and outcome of medical visits. Therefore, gender-specific differences in response to cost containment may also occur. OBJECTIVE: The purpose of this study was to assess the effect of physician gender on changes in prescribing patterns of angiotensin-converting enzyme (ACE) inhibitors after the implementation of reference pricing for prescription drugs in British Columbia, Canada. METHODS: Reference pricing is a cost-sharing policy by which use of high-priced medication requires out-of-pocket payment of the price difference between the cost-sharing drug and a lower-cost drug within the same class. In British Columbia, reference pricing for ACE inhibitors was introduced on January 1, 1997. Analysis was carried out on linked pharmacy and medical service claims data on 927 female and 2922 male physicians treating 47,680 Pharmacare Plan A enrollees who were aged >-65 years and were prescribed a high-priced ACE inhibitors before the implementation of reference pricing. RESULTS: Female physicians (24.1% of all physicians) were younger, treated more female patients, had patients with fewer chronic illnesses, and worked more often as general practitioners than did male physicians. The patients of female physicians were more likely to receive a written physician-requested exemption from copayment, according to a multivariate logistic regression analysis (odds ratio [OR], 1.25; 95% CI, 1.04-1.50). Data suggested that patients of female physicians were more likely to stop antihypertensive drug therapy (OR, 1.43; 95% CI, 0.96-2.13); however, this was independent of the new copayment policy. CONCLUSIONS: The results provide empirical evidence that physician gender is associated with slightly different patient management strategies regarding physician-requested exemptions after the start of a new drug cost-sharing policy. However, these differences are unlikely to have meaningful clinical or economic consequences.

Health measures: differentiating associations with gender and socio-economic status.

BACKGROUND: In a cohort survey on health-related lifestyles, four different measures of health were analysed with regard to their associations with gender, socio-economic and psychosocial factors. METHODS: The survey was carried out in Berne, Switzerland. Response rate was 64% in the initial interview and 83% in the second interview, from which the data presented were derived, resulting in 923 participants aged 56 to 66 years. Along with socio-economic and psychosocial parameters, four self-report health measures were obtained, namely self-rated health, physical fitness, number of medical conditions and restrictions caused by medical conditions. Regression analysis was used to investigate and compare their associations with gender, socio-economic and psychosocial factors and relevant interaction terms. RESULTS: Gender was statistically significantly associated with physical fitness, number of medical conditions and subsequent restrictions. Education and income showed statistically significant associations with self-rated health and fitness. Psychological factors were statistically significantly associated with all health measures. Gender showed to interact with education, income interacted with internal health locus of control. Analyses with separated genders showed that the association of socio-economic status with self-rated health and fitness was statistically significant in women only. CONCLUSION: The different health measures showed considerable variation in strengths of association with health-related factors, most noticeably so with gender and socio-economic status. The choice of health measures in population studies should comply with the intention to analyse its associations with any of those related factors, or, in reverse, with the wish to prevent their confounding properties.