Table 2 Open in new tab Strengths and...

Strengths

Weaknesses

Regulatory sponsored studies

Arrives early after marketing

Patient selection may not be representative

Targeted data collection

Learned society academic studies

Targeted data collection

Patient selection need not be representative

Usually wide geographical representation

Quality of outcome registration can vary

Nationwide or regional registries

Large scale

Data quality may be limited given use of clinical documentation

Less bias in patient selection

International generalizability uncertain

Low cost

Claims data

Complete selection of data within an administrative unit

Many clinically important data (both independent and outcome variables) may not be available

Low cost

Quality of data may be limited

Investigator-initiated and industry-sponsored studies

Multiple centres

Reimbursement for participation can influence patients who consent to intervention

Careful monitoring of data collected

Centre selection can result in unrepresentative patients

Targeted data collection

Questions may be designed to ensure a higher probability of a favourable outcome

Hospital cohorts

Uniform patient selection

Patient selection not representative

Similar expertise to all patients

Data quality may not be high

Expertise of selected centres may not be generalized

Table 2

Strengths and weaknesses of common data sources

Strengths	Weaknesses
Regulatory sponsored studies
Arrives early after marketing	Patient selection may not be representative
Targeted data collection
Learned society academic studies
Targeted data collection	Patient selection need not be representative
Usually wide geographical representation	Quality of outcome registration can vary
Nationwide or regional registries
Large scale	Data quality may be limited given use of clinical documentation
Less bias in patient selection	International generalizability uncertain
Low cost
Claims data
Complete selection of data within an administrative unit	Many clinically important data (both independent and outcome variables) may not be available
Low cost	Quality of data may be limited
Investigator-initiated and industry-sponsored studies
Multiple centres	Reimbursement for participation can influence patients who consent to intervention
Careful monitoring of data collected	Centre selection can result in unrepresentative patients
Targeted data collection	Questions may be designed to ensure a higher probability of a favourable outcome
Hospital cohorts
Uniform patient selection	Patient selection not representative
Similar expertise to all patients	Data quality may not be high
	Expertise of selected centres may not be generalized

Strengths	Weaknesses
Regulatory sponsored studies
Arrives early after marketing	Patient selection may not be representative
Targeted data collection
Learned society academic studies
Targeted data collection	Patient selection need not be representative
Usually wide geographical representation	Quality of outcome registration can vary
Nationwide or regional registries
Large scale	Data quality may be limited given use of clinical documentation
Less bias in patient selection	International generalizability uncertain
Low cost
Claims data
Complete selection of data within an administrative unit	Many clinically important data (both independent and outcome variables) may not be available
Low cost	Quality of data may be limited
Investigator-initiated and industry-sponsored studies
Multiple centres	Reimbursement for participation can influence patients who consent to intervention
Careful monitoring of data collected	Centre selection can result in unrepresentative patients
Targeted data collection	Questions may be designed to ensure a higher probability of a favourable outcome
Hospital cohorts
Uniform patient selection	Patient selection not representative
Similar expertise to all patients	Data quality may not be high
	Expertise of selected centres may not be generalized