Designing a Flexible Study

If your trial is flexible—meaning that the protocol calls for decision points that determine which Intervals (groups of visits) and DCIs (equivalent to CRFs) are expected for each particular patient—use the Enhanced DCI Book feature to define the decision points and rules that determine Interval and DCI expectedness for each patient as he or she progresses through the trial.

Implementing a flexible study design in Oracle Clinical includes the following:

Schedule - Define the study schedule, including Intervals (Phases, Periods, and Subperiods) and visits, or Clinical Planned Events (CPEs), in the same way as for nonflexible studies in Oracle Clinical, associating CPEs with Intervals and defining both CPEs and Intervals in chronological order. In the DCI Book, associate DCIs with CPEs.

Rules - Create a conditional schedule by defining rules. Identify the particular Questions and responses to them (the rule trigger) that determine which DCIs or Intervals (the rule target) are expected for a particular patient.

An Intervals or DCI that is the target of a rule is conditional; it is expected for a particular patient only if the data condition specified in the rule trigger is met for that patient. An Interval or DCI that is not the target of a rule—or dependent on another Interval or DCI that is the target of a rule—is unconditional. It is expected for all patients assigned to the DCI Book.

It is also possible to define a rule with an Interval as a target and an action called Bypass To; when the trigger condition is met for this type of rule the target Interval becomes expected for the patient and all intervening Intervals are skipped; they are no longer expected.

Calculating Data Expectedness for Patients - As sites enter data for a particular patient, Oracle Clinical recognizes the decision points you have defined as triggers in the DCI Book's rules, determines whether the value entered satisfies the defined condition and if so, makes the DCI(s) or Interval(s) defined as the target of that rule expected for that patient.

CPE and CRF Display in RDC Onsite - RDC Onsite displays only CRFs and CPEs that are either collected or expected for each particular patient. They may be expected because the corresponding DCIs and Intervals are defined as unconditional or because their condition has been met.

Whenever a user updates data that triggers a rule, the system recalculates expectedness for the patient and RDC Onsite adjusts the display of expected CPEs and CRFs as necessary. If a CRF that already has data collected becomes not expected due to a data update, RDC Onsite still displays the CRF icon but an "N" for not expected appears next to the CRF icon.

Note:

See also Flexible Study Design Examples

For more information, see:

• Enabling Flexible Study Functionality
  
• Defining Intervals and CPEs for Flexible Studies
  
• Using a Single DCI Book
  
• Using Rules
  

Enabling Flexible Study Functionality

To take advantage of flexible study functionality in the Enhanced DCI Book window and in RDC Onsite, you must first define the study as flexible. When you create a new study using Easy Study Design, the system prompts you for the Flexible setting.

You can also make an existing study flexible in the Clinical Study States window under Conduct, Security if no test or production data has been entered.

Note:

If the study has an active DCI Book, you must set it to Provisional and back to Active before you can use it for data entry. If you have patients assigned to the book and you plan to use the book with RDC Onsite, you must also run the Calculate Expectedness job.

You cannot reverse the Flexible setting from checked to unchecked if the study has production data.

Defining Intervals and CPEs for Flexible Studies

As you plan your study schedule and how to define it, consider the ramifications for flexible studies:

• Interval and CPE order determine the default order of the DCI Book. You can define rules that allow a particular patient to skip certain Intervals—for example, to follow Treatment Arm A instead of Treatment Arm B—but you cannot use the DCI Book to change Interval or CPE order. A patient can only move from one Interval or CPE to another defined later in the study schedule. For example, if a rest period is required after Treatment A and after Treatment B, you must define it after both Treatment A and Treatment B in the study schedule; see complex-examples.html#GUID-5AB82201-267C-4541-9073-0572EB42F4D1__CHDBHGFD for an example. See Defining a Study Schedule and Creating Clinical Planned Events for information on defining Intervals and CPEs.

• Using rules, you can make a whole Oracle Clinical Interval (Phase, Period, or Subperiod) that directly contains CPEs, expected (or skipped) for a particular patient if that patient's data meets conditions you define. Therefore define groups of CPEs that you want to enable or skip all at once as a single Interval.

• You can make a particular DCI expected for a particular patient if that patient's data meets conditions you define. Even if the Interval as a whole is expected for a patient, you can define rules in such a way that every DCI at every CPE is not necessarily required.

• You may want to give Intervals and CPEs nonnumeric names so that you can incorporate these names, or abbreviations of them, in DCI Book page numbers. In a DCI Book, each DCI, or Book page, must have a unique number. Normally no patient will have data collected for every DCI, so if you use purely numeric page numbers there will be gaps; see Sample Numbering Schemes for more information.

  Note:

  When you create a study using Easy Study Design, the system automatically creates a Phase Interval called DEFAULT STUDY PHASE. If you are creating a flexible study you cannot use this default Interval and must explicitly create Intervals.

Using a Single DCI Book

Design flexible studies with only one DCI Book. You can define multiple treatment arms and many decision points and rules within a single DCI Book. Use the same DCI Book for all patients in study.

The only time you may want to have more than one DCI Book for a study is when the protocol is amended during the trial. In that case, you may want to copy the DCI Book, make the necessary changes, and reassign patients to the revised Book.

Using multiple DCI Books that include different portions of the study schedule is not recommended because there is no validation that the definitions of the two Books are consistent with each other. Migration between such unrelated books will be successful but the new expectedness calculation will reflect only the structure and rules of the new Book.

Using Rules

Each rule consists of a trigger—the data condition that must be met to trigger the rule, and a target—the DCI or Interval that is enabled when the condition is met.

To determine what rules your study requires, it may help to draw a diagram similar to those in Flexible Study Design Examples

Validation Checks covers information about rule definition restrictions.

For more information, see:

• Rule Triggers
  
• Rule Targets
  

Rule Triggers

In most cases, a trigger is one or more specified values that, when entered in response to a specified Question in a specified DCI at a specified CPE, constitute the condition that invokes the rule and makes the trigger DCI or Interval expected for the patient. It is also possible to specify that any data collected for a specified DCM constitutes the condition; see "Using the Any Data Trigger" under Rule Triggers.

Questions used as a rule trigger must:

• have a Discrete Value Group (DVG) of type Internal or Thesaurus. The user must select the Question response value from the list of valid values specified in the DVG; see Creating and Using DVGs.

• be of type CHAR

• not be a complex Question

• not be included in a repeating Question Group

Question response values cannot be numeric and they cannot be a range of numbers. However, if you need a numeric value or range of values to serve as a trigger, you can effectively do that using a Derivation Procedure.

Using Derivation Procedures

To convert a numeric value to a character value that can be used as a trigger, define a numeric Question to collect the numeric data, a derived Question to hold the character value, and a Derivation Procedure that takes the numeric Question response value as input, performs a calculation, generates a character value, and supplies the character value as the response to the derived character Question.

For example, if you want to require additional tests for patients whose systolic blood pressure is over 160, you can:

• Define numeric Question SBP to collect the patient's actual systolic blood pressure.

• Define a Derivation Procedure to take the value of SBP and determine whether it is greater than, equal to, or less than 160 and, if it is greater than 160, to return a value of Y and if not, to return a value of N.

• Define derived Question DSBP to hold the derived value—either Y or N. (You must define the derived Question before you define the Derivation Procedure.)

• Assign a Discrete Value Group (DVG) with the values Y and N to Question DSBP.

• Define the rule trigger to be the value of DSBP, in a specified DCI and CPE, equal to Y.

• Define one or more DCIs that collect the additional data—at one or more CPEs—as the target of the rule.

If a patient's systolic blood pressure has a value greater than 160, the Derivation Procedure generates a value of Y, the rule is triggered, and the specified DCIs become expected for that patient and visible in RDC Onsite.

The timing of the execution of the Derivation Procedure depends on the setting of its Exec Content field; if it is set to ON-LINE/DCM the Procedure is executed when the RDC data entry operator saves the value of Question SBP (if the RDCI status is high enough). If the Procedure generates a value of Y, the target DCI is immediately evaluated as expected for the patient and displayed in RDC Onsite.

See Validation and Derivation Procedures for more information on defining Derivation Procedures.

Note:

The RDCI (collected data) must have a status of Batch Loaded or Pass 1 Complete or beyond. This might not be the case if the RDC user used the Save rather than the Save Complete action, or if data was entered through Oracle Clinical and Pass 2 was required for the study.

Data entered through Oracle Clinical or earlier versions of RDC Onsite triggers rules and makes Intervals and DCIs expected. However, expectedness is not interpreted by the user interface except in RDC Onsite.

Using the Any Data Trigger

The Any Data trigger satisfies the rule condition if the trigger DCI has any entered data at all. You do not need to define a particular Question as the rule trigger. As soon as any data is entered and saved for the DCI for a patient (if the RDCI has a status of Batch Data Loaded or Pass 1 Complete or beyond), the target Interval is enabled, or expected, for that patient.

For example, if the Rest or Followup Interval is not the target of any other rule, but must come after each treatment Interval, use Any Data in a DCI in the treatment Interval as the trigger to make the Rest or Followup Interval expected for a patient.

If you define a Rest or Followup Interval after each treatment Interval in the study schedule, it will be expected even without being the target of a rule. However, by defining the rule, the target Interval does not appear in RDC Onsite until they are logically expected.

Any Data is available only for Interval rules, and only with an action of Enable.

Rule Targets

A rule target can be one or more DCIs or Intervals. Rules that have one or more DCIs as the target are called DCI Rules. rules that have one or more Intervals as the target are called Interval Rules.

DCI Rules

In a flexible study, all patients may not have the same assessments at every CPE. For example, a subset of pages may be required only for patients assigned to a specific Treatment Group, or detailed information may be required only if certain criteria are met, such as a detailed smoking history for smokers. Use DCI rules to make this possible.

For example, to make the detailed smoking history expected only for patients who smoke, define a DCI to contain the smoking history Questions, associate the DCI with the appropriate CPE(s) in the DCI Book Pages window, and define a DCI rule with the DCI as the target.

There are two types of DCI rules:

• Within CPE: The target DCI is in the same CPE as the trigger DCI, and is enabled in that CPE when the rule is triggered. (The same target DCI may be included in other CPEs, but other occurrences of the DCI do not become expected for a patient when the "Within CPE" DCI rule is triggered.)

• Across CPEs: The target DCI is enabled in every CPE where it exists, including the same CPE as the trigger DCI if it exists there, when the rule is triggered.

Note:

If only a few Questions need to be conditional on a particular Question response, you may prefer to use conditional branching instead of a "Within CPE" DCI rule; see Defining Conditional Branching.

A DCI can be the target of only one rule. A DCI that is the target of a rule can also be the trigger for a rule, but it can act as a trigger only when it is expected and collected for a patient. See Defining DCI Rules for further information.

Interval Rules

In a flexible study there are multiple pathways through the trial for different patients. There may be entire Intervals (groups of CPEs) that are not required for some treatment groups, but are for others.

An Interval can be defined in the Oracle Clinical study schedule as a Phase, Period or Subperiod. Phases can contain Periods, and Periods can contain Subperiods. You can define Phases, Periods, and Subperiods as targets, as long as the rules do not conflict with each other (see Validation Checks).

An Interval can be the target of multiple Interval rules. It becomes expected for the patient the first time a rule targeting it is triggered.

If you do not define Interval rules, all Intervals are expected for all patients, though DCIs within Intervals may be conditional if they are the target of DCI rules. See Defining Interval Rules for further information.

There are two types of Interval rules:

• Enable: The target Interval(s), with all their unconditional CPEs and DCIs, become expected for the patient when the trigger condition is met. You can use an Enable Interval rule, for example, to make Phase I of a particular treatment arm expected, depending on the response to a particular Question; see basic-examples.html#GUID-D1C3C36D-DC35-48A6-94D0-88FFF1BFA4FC__CHDFGFCA.

• Bypass To: The target Interval becomes expected for the patient when the trigger condition is met and all intervening Intervals—all Intervals that occur between the trigger and target Interval—are not expected. Any remaining expected DCIs in the trigger Interval are still expected.

  You can use Bypass To Interval rules, for example, to handle patients who discontinue prematurely from the trial, using a Question as straightforward as "Will the patient continue in the trial?"; see basic-examples.html#GUID-D63A9F67-C8CB-4731-9682-3BFAC55E014C__CHDDHFED.

  A Bypass To Interval rule can have only one Interval as a target, and that target cannot be Next Interval. Bypass rules take precedence over Enable rules of all types (DCI—Within and Across—and Interval) if there is a conflict.

In Enable Interval rules, you can specify Next Interval as the target instead of a named Interval. Next Interval is defined as the Interval containing the next CPE in ascending CPE number order that is in an Interval different from the one that contains the trigger DCI.

Using Next Interval has the following advantages:

• It reduces the number of rules required when the same trigger condition may appear repeatedly and the target is always the next Interval; for example, in oncology studies where patients continue on cycles of treatment until the disease progresses or they respond to treatment. You can define a single rule with a Next Interval target to move each continuing patient from one cycle to the next. See basic-examples.html#GUID-34B98ABB-CB82-4DE9-AF46-C8F1ACD60D57__CHDEHGEF for an example.

• All potential Next Intervals—all Intervals whose preceding Interval contains a the trigger of a Next Interval rule—are treated as conditional by the system; RDC Onsite does not display for a particular patient them unless they become expected.

The target Interval may be an Interval within the trigger Interval; for example, if the trigger Interval is a Phase that contains CPEs directly and also contains Periods, a DCI in a CPE contained directly in the Phase could trigger the next Period in the same Phase.