Multi-Site Analysis for Independent Data Sources

The multi-site analyses included in this suite are intended to be executed against data that are all stored in the same place. However, there may be some instances where the data associated with each site is stored in independent locations. This vignette outlines how the multi-site analysis can be executed in these instances.

After following the instructions to reproduce the analysis, you will also need to change the output_function column to tell the qvd_output function which check you executed. Reference the table below for the labels that are associated with each check:

Check Type	output_function
Multi Site, Exploratory, Cross-Sectional	qvd_ms_exp_cs
Multi Site, Exploratory, Longitudinal	qvd_ms_exp_la
Multi Site, Anomaly Detection, Cross-Sectional	qvd_ms_anom_cs
Multi Site, Anomaly Detection, Longitudinal	qvd_ms_anom_la

Multi Site Exploratory Analysis

The process for the exploratory analysis is the same for both the cross-sectional and longitudinal configurations.

First, execute either of the Single Site, Exploratory analyses, configured appropriately for your study, against each data source.

library(quantvariabledistribution)

my_table <- qvd_process(cohort = my_cohort,
                        multi_or_single_site = 'single',
                        anomaly_or_exploratory = 'exploratory',
                        time = T / F,
                        ...)

Then, combine these results into a single table with the different sites delineated in the site column.

my_final_results <- my_table1 %>% dplyr::union(my_table2) ... %>%
  dplyr::union(my_table_n) %>%
  dplyr::mutate(output_function = '{see table above}')

Multi Site Anomaly Detection

For anomaly detection analysis, start by executing the same steps as the exploratory analysis. Then, you will execute the relevant anomaly detection algorithm against the resulting table. See below for the different processes for cross-sectional and longitudinal analysis.

Cross-Sectional

For a cross-sectional analysis, the compute_kl_divergence function in this package should be executed against your results. This will compute the Kullback-Leibler divergence of each site’s value distribution to the overall value distribution.

The kl_log_base value can be selected by the user out of the following options: log, log2, or log10. The standard default is log2.

df <- 
  quantvariabledistribution:::compute_kl_divergence(frequency_tbl = my_table,
                                                    kl_log_base = kl_log_base) %>%
  dplyr::mutate(output_function = '{see table above}')

Longitudinal

For a longitudinal analysis, the qvd_euclidean function in this package should be executed against your results. This will compute the Euclidean distance from the site’s time series to the overall time series.

The euclidean_stat value can be either mean or median, as desired by the user.

df <- 
  quantvariabledistribution:::qvd_euclidean(fot_input_tbl = my_table,
                                            grp_vars = c('site', 'value_type'),
                                            euclidean_stat = euclidean_stat) %>%
  dplyr::mutate(output_function = '{see table above}')