Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one

Home/Insights/Evidence synthesis/Integrating health system data with systematic reviews

Integrating health system data with systematic reviews

Reviews tell you what worked in trials; health system data tell you what is happening in your population. Each answers what the other cannot.

Tools and data repositories

Why combine them

A systematic review establishes efficacy or effectiveness under the conditions studied. It does not tell a health system what its own population looks like, how many patients would be eligible, what current practice is, or what the local baseline risk is.

Routinely collected data — electronic health records, claims, registries, audit data — answer those questions and cannot establish effect. Together they support a decision; separately neither does.

Where local data change the interpretation

Baseline risk. A relative risk reduction from a review translates into very different absolute benefit depending on local event rates. The number needed to treat is a local quantity.

Applicability. Comparing the review’s included populations with the local population shows whether the evidence is indirect, which is a GRADE domain.

Current practice. A review’s comparator may not be what is done locally, which changes the relevant comparison entirely.

Eligibility and volume. How many local patients meet the trial criteria determines whether an effective intervention is worth implementing.

The cautions

Routine data were collected for administration and billing, so coding reflects reimbursement rules as much as clinical reality, and a diagnostic code is a proxy rather than a diagnosis. Completeness varies by site and over time, and a coding change looks exactly like a change in incidence.

Combining the two does not license a causal claim from the routine data. Local data describe; the review supplies the effect estimate.

Analyses of routinely collected data should be reported to RECORD, which extends STROBE for this purpose, with code lists published.

Doing it transparently

State the data source, version and extraction date; publish the code lists; separate clearly what comes from the review and what from local data; and present absolute effects for the local population alongside the review’s relative ones.

In practice

Fix, flag or query.

The distinction that keeps an edit defensible: what an editor may change silently, what must be shown to the author, and what only the author can answer.

  • Use local data for: baseline risk, eligibility and current practice — not effect.
  • Publish: the code lists defining local cohorts.
  • Report routine-data analyses to RECORD.

Sources

  • RECORD Statement, for routinely collected health data
  • GRADE Working Group, on indirectness and applicability
  • AHRQ EPC Methods Guide, on applicability

Sources are given so each claim can be checked at its origin. Where a tool’s current maintenance status could not be confirmed, the page says to check it rather than asserting it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.