Table 1 Features of trials appropriate for clinically-integrated versus more traditional randomized trials.

Tests, procedures, questionnaires

All data needed to address the study question would be of value for the clinician during routine follow-up. All tests, procedures and questionnaires would be given to patients irrespective of participation.

Example: A test for cancer recurrence.

Some data required to answer study-specific questions would not be taken during routine care.

Example: blood draw for a molecular marker thought to predict response to treatment.

Treatments

Patients very unlikely to have strong preferences for one or other treatment.

Example: Two alternative suturing techniques during surgery.

Many patients may have a strong preference for one or other treatment.

Example: Radical versus breast conserving surgery.

Comparisons

Can only compare two active treatments.

Example: Two widely used anti-depressants of proven value.

May compare an active treatment to placebo or no treatment control.

Example: Novel anti-depressant versus placebo.

Patients

Most patients are randomized.

Only a proportion of patients are randomized.

Eligibility criteria

Eligibility criteria should be minimized.

Example: all patients undergoing radical prostatectomy are eligible.

Eligibility criteria can be restrictive.

Example: restrictions on comorbidities in a trial of a novel drug.

Information technology

Trial depends heavily on information technology.

Trial can be lo-tech.