1.4 Summary Tables in Descriptive Studies of Background Rates or Exposures with Follow-Up Time

In studies of background rates as well as studies of exposures with follow-up time, the QRP Reporting Tool outputs tables on the incidence and/or prevalence of medical product use or an event of interest over a specified time period. In studies of exposures with follow-up time or concomitant use, these tables additionally include time at risk and the rate of events during exposure. Summary results are presented overall, as well as by user-specified strata.

1.4.1 Output

Using parameters in QRP Report Table File, users can specify which tables with or without stratifications to output in the report. Users can request stratifications on the following factors: year, sex, age group, race, Hispanic ethnicity, month, quarter, geographic region variables, user-defined characteristics, as well as various combinations of the above factors. Users can leverage the template QRP Report Table Columns File located within the ‘TemplateFiles’ folder of the QRP Reporting Tool file structure to specify which columns to include in the table(s). There are two templates available: one showing default columns and one showing all columns that can be calculated using the default columns.

Table 1.5 portrays the columns available for output in background rate and exposures with follow-up time summary tables, with columns marked with an asterisk representing the default columns and “per X” representing the multiplier specified by the user in the COLUMN parameter of QRP Report Table Columns File. Users should note that the column name and label pre-specified in the template files may not necessarily represent the user’s study design (for instance, “New Users” may not pertain to a Study of Background Rates where no medical product of interest is specified or where prevalent use is required), and should consult the technical documentation for the parameters used to calculate each column to understand more.

Table 1.5: Columns Available in Summary Tables for Descriptive Studies of Background Rates and Exposures with Follow-Up Time
  Background Rates Exposures with Follow-up Time Concomitant Exposures with Follow-up Time
Patients
New Users1 X X X
New Users per (X) Eligible Member(s) X X
New Users per (X) Eligible Member-Days X
Episodes
New Episodes X X X
Events
New Episodes with an Event1 X X
Total Number of Events X X
Episodes with an Event per User X X
Episodes with an Event per (X) Episodes X X
Episodes with an Event per (X) Eligible Members X
Episodes with an Event per (X) Member-Year(s) at Risk X X
Event Rate per (X) Person-Day(s) X X
Event Rate per (X) Person-Month(s) X X
Event Rate per (X) Person-Year(s) X X
Study Population
Eligible Members1 X X
Eligible Member-Days1 X X
Eligible Member-Years1 X X
Dispensing Information2
Adjusted Dispensings1 X X X
Raw Dispensings1 X X X
Days Supply1 X X X
Amount Supplied1 X X X
Adjusted Dispensings per Users X X X
Raw Dispensings per User X X X
Days Supply per User X X X
Days Supply per Episode X X X
Amount Supplied per Episode X X X
Days Supply per Adjusted/Raw Dispensings X X X
Study Time
Amount Supplied per Adjusted/Raw Dispensings X X X
Time to Data End (Days) X X
Days at Risk1 X X
Years at Risk1 X X
1 Indicates this column is output by default when using the "Default" Table Columns File
2 Dispensings indicate the number of prescription dispensings that were used in index date-defining records. When multiple dispensings in the same STOCKGROUP occur on the same day, the "raw" dispensings metric counts each whereas the "adjusted" dispensings metric counts them as one.

If Eligible Members, Eligible Member-Days, and/or Eligible Member-Years columns are output, then the following footnote is printed beneath the table: “[Column Label] are reflective of the number of patients that met all cohort entry criteria on at least one day during the query period” where [Column Label] represents the label of the relevant column(s).

If data are stratified by race, then the following footnote is output: “Race data may not be completely populated at all Data Partners; therefore, data about race may be incomplete.” Additionally, if race categories with ten or fewer counts of patients are collapsed into the “Unknown” category, the following footnote is output: “Includes members classified as having an unknown race by the Data Partner and patients in race categories where the total member count is between one and ten.”

1.4.2 User Options

Summary tables are titled ‘Summary of [Report Title] in the [Database Name] from [Query Start] to [Query End], by [Stratifying Variable]’ where:

  • [ReportTitle] is user-specified in QRP Report Label File
  • [Database Name] is user-specified in QRP Report DP Info File]
  • [Query Start] is user-specified in qrp.MonitoringFile
  • [Query End] is the maximum date of all the included Data Partner end dates (unless an explicit follow-up end date is specified in QRP Report Monitoring File and is earlier than the maximum Data Partner end date, in which case the end date is the follow-up end date)
  • [Stratifying Variable] is user-specified in QRP Report Table File. When no stratifying variable is requested, the title ends after [Query End].

Summary tables output one row for each cohort group. If desired, users can choose to output multiple cohort groups under a combined header by specifying the appropriate parameter in QRP Report Label File.

Finally, users can relabel and reorder columns, as well as specify the number of decimal places to display, using the QRP Report Table Columns File.

1.4.3 Calculations

If the “New Users per (X) Eligible Member(s)” or the “Episodes with an Event per (X) Episodes” columns are output, then the following formula is used to compute the confidence interval:

The large sample, normal theory \(\left(1-\alpha\right)\times100%\) confidence interval for the proportion \(\hat{p}\) is given by

\[\begin{equation} \hat{p}\pm Z_{(1-\frac{\alpha}{2})}\sqrt{\frac{\hat{p}\left(1-\hat{p}\right)}{n}} \tag{1.12} \end{equation}\]

where \(\hat{p}=\frac{events}{n}, n=\) the sample size, and \(Z_{(1-\frac{\alpha}{2})}\) is the \(\left(1-\frac{\alpha}{2}\right)\times100th\) percentile of the standard normal distribution. For a 95% confidence interval, \(Z_{(1-\frac{\alpha}{2})}\) is commonly approximated as 1.96.

If the “New User Rate per (X) Eligible Member-Days” or the “Event Rate per X Person-Days/Months/Years” columns are output, then the following formula is used to compute the confidence interval:

The large sample, normal theory \(\left(1-\alpha\right)\times100%\) confidence interval for the proportion \(\hat{\lambda}\) is given by

\[\begin{equation} \exp{\left\{log{\left(\hat{\lambda}\right)}\pm Z_{1-\frac{\alpha}{2}}\sqrt{\frac{1}{events}}\right\}} \tag{1.13} \end{equation}\]

where \(\hat{\lambda}=\frac{events}{person-time}\), and \(Z_{(1-\frac{\alpha}{2})}\) is the \(\left(1-\frac{\alpha}{2}\right)\times100th\) percentile of the standard normal distribution. For a 95% confidence interval, \(Z_{(1-\frac{\alpha}{2})}\) is commonly approximated as 1.96.