Question 1 (18 marks)
Read the paper Moy, F. M., Darus, A., & Hairi, N. N. (2015). Predictors of handgrip strength among adults of a rural community in Malaysia. Asia Pacific Journal of Public Health, 27(2), 176-184.
Critically appraise of the statistical material in this paper against items 10, 12-17 of the STROBE checklist. Present your review as a 400-500 word (approx.) report.
Question 2 (22 marks)
Using R Commander and the data set from the sample of grandparents attending play groups with their infant grandchildren in Parramatta assigned to you, address the following research questions:
Does average hand grip strength (variable ‘grip’) differ between grandparents whose occupational history is coded ‘heavy manual’ and those whose occupational history is coded as ‘other’ (variable ‘occ’) after correction for body mass index (variable ‘bmi’), in the population of grandparents attending play groups in Parramatta? (You should address this question using linear regression and include associated descriptive analyses.)
Note: To answer this question, you need to use R Commander and the data set assigned to you for assignments. This data set contains the (fictitious) data from a random sample of grandparent carers of infants in Parramatta and is the same data set as that you have previously used in Assignments 1 and 2. See ‘Description of your data set.docx’ for the descriptions of the variables.
Note: This assignment is assessing your skills, not the skills of the computer. You will need to include graphs from R Commander into your assignment but all other R Commander output will attract 0 marks and is discouraged. It is your task to identify the relevant results in the R Commander output and write these up in your assignment.
401077 Introduction to Biostatistics, Spring 2019
Provide your appraisal of the strengths and weaknesses of the presentation of the statistical material in Moy et al (2015) against items 10, 12-17 of STROBE (about 400-500 words, 18 marks)
Strobe 10: Sample Size.
In Moy et al (2015) the sample size is provided in the method section. Clarity on the sample size and standard error is provided by the author but there is no mention of the error margin and how the sample size was calculated. The author also mentions that the total households selected from the 5 districts were 1250.
Strobe 12: Statistical Methods.
Describe all statistical methods, including those used to control for confounding.
The author moderately explains the statistical methods. In the last paragraph of Method section, the author has stated that all continuous variables were analysed using t-test or 1-way analysis of variance. Categorical variables were analysed using chi-square test. All statistical analyses were satisfied by gender as evidence indicates gender difference in handgrip strength. Backward multiple regression analyses were carried out to assess possible predictors of hand grip strength. Additionally, the authors mention about the inclusion of all analysis variables and the regression variables and the regression excluded imputed outcomes, but they have not clearly mentioned about the procedures, data transformation, calculations of the confidence interval or the statistical software used to report the result.
Describe any methods used to examine subgroups and interactions.
Moy et al (2015) completely complies with the STROBE 12(b). In the method section it is clearly stated that the main aim of this study was to explore the rural community with a wide spectrum of occupation from services, agriculture and manufacturing. Out of 10 districts, 5 districts were randomly selected. From each district one third of the total households were randomly selected. The total households selected from the 5 districts were 1250.
Explain how missing data were addressed.
The author fails to mention how the missing data was treated.
Cross-sectional study: If applicable, describe analytical methods taking account of sampling strategy.
The authors clearly mention about the sampling strategies in the Methods section. Therefore, the authors completely comply with the STROBE 12 (d).
Describe any sensitivity analyses
No sensitivity analyses are mentioned.
Strobe 13: Participants
Report number of individuals at each stage of study – e.g. numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up and analyses.
The authors have mentioned about the numbers of participants in the methods section and about the eligibility. But, number of participants of the sub-groups are not mentioned. Thus, the STROBE 13 (a) is not completely addressed by the authors.
Give reasons for non-participant at each stage.
The authors have not specified for any reasons for the non-participation and the missing data. Thus, the author does not comply to STROBE 13(b).
Consider use of a flow diagram.
As the study is not a complex observational study there is no need for a flow diagram. There for no compliance with STROBE 13 (c).
Strobe 14: Descriptive data
Give characteristics of the study participants (e.g. demographic, clinical, social) and information on exposure and potential confounders.
Moy et all (2015) has mentioned about the confounders in Covariates section and in Table 1 also. Thus, the study completely complies with the STROBE 14 (a).
Indicate number of participants with missing data for each variable of interest.
Moy et al (2015) in the Method section has mentioned about the accountability for the missing data but fail to provide the number for the missing data for all the interested variables. Therefore, no compliance with STROBE 14 (b).
Cohort study: Summarise follow-up time – e.g. average and total amount.
In table 1 and table 3 summarises the same type of information on prevalent outcome events or summary measures. Therefore, compliance with the STROBE 14 (d).
Cross-sectional study—Report numbers of outcome events or summary measures.
The study compliances with the STROBE 15 as the authors summaries about the outcome and report it in the Table 1, 2 and 3 for the mean, association coefficient and the confidence intervals.
Strobe 16: Main Results
Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (e.g. 95% confidence interval). Make clear which confounders were adjusted for and why they were included.
The study compliances with STROBE 16 as the authors clearly mentioned about the confounders in the Method section. The adjusted and unadjusted confounders are also mentioned additionally in the Tables 1, 2 and 3 along with the result section.
Report category boundaries when continuous variables were categorised.
Moy et al (2015) states that the occupation history was evaluated based on variables such as “ever worked”, “current working status”, “job groups”. But the author fails to mention the range. Thus, the study does not comply with STROBE 16(b).
If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period.
Other analyses: Report other analyses done—e.g., analyses of subgroups and interactions and sensitivity analyses.
Compliance with STROBE 17 as the Confidence interval for the subgroups is estimated in the method section and in the Tables 1, 2 and 3.
Question 2: Note: Students will get different answers as the data sets differ.
Present the findings of your descriptive analyses (graphs, tables and about 100-150 words, 8 marks)
The same is show in the graphs above. First graph is the occupation of grandparents where 1 is heavy manual and 0 is other. Second graph is the grip strength graph where we can see that it is a normally distributed graph.
Present the findings of relevant regression models and inferential analyses (about 150-200 words, 10 marks)
lm(formula = data$grip ~ data$occ + data$bmi, data = data)
Min 1Q Median 3Q Max
-9.5016 -2.4935 0.2917 2.5799 7.5651
Estimate Std. Error t value Pr(>|t|)
(Intercept) 41.71217 2.00956 20.757 < 2e-16 ***
data$occother 0.68155 0.52259 1.304 0.193
data$bmi -0.36898 0.06962 -5.300 2.71e-07 ***
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 3.707 on 230 degrees of freedom
Multiple R-squared: 0.1123,Adjusted R-squared: 0.1046
F-statistic: 14.55 on 2 and 230 DF, p-value: 1.127e-06
Grip = 41.71217 + 0.68155 x Occupation – 0.36898
P-value should be less than 0.05 which is only with bmi and not occupation which means bmi is making a significant difference and occupation is not significantly impacting the model.
Residual standard error is 3.707 which means the error is estimated as the +- 3.707 in this model. Multiplied and Adjusted R-Square is almost same 0.1123 and 0.1046 respectively, which means the fit of the model is not at all good i.e. independent variables are only explaining 10% of the data and rest of the data is not being explained. R-Square also tells how fit the model is to be considered. Here the fit is low so we can say that the data is not good to predict about the grip.
Provide your answer to the research question (40-80 words, 4 marks)