Home > DARE Reviews > The effect of feedback to general...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews [Internet]. York (UK): Centre for Reviews and Dissemination (UK); 1995-.

Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews [Internet].

The effect of feedback to general practitioners on quality of care for people with type 2 diabetes: a systematic review of the literature

TL Guldberg, T Lauritzen, JK Kristensen, and P Vedsted.

Review published: 2009.

CRD summary

This review concluded that feedback to general practitioners on diabetes care appeared a promising tool for quality improvement in care of patients with Type 2 diabetes and further research was required (especially to assess electronic feedback). The authors' conclusions should be interpreted with caution given potential for missing data, variation between studies and an unknown risk of bias.

Authors' objectives

To assess the effects on quality of care of providing feedback to general practitioners (GPs) on diabetes care for patients with Type 2 diabetes.


Medline, The Cochrane Library and Embase were searched up to October 2008 for studies published in English. Despite this limitation one unpublished study was considered for inclusion. Search terms were reported. Reference lists of retrieved papers and the related articles facility in PubMed were used to check for further studies.

Study selection

Randomised controlled trials (RCTs) that assessed the effects of providing feedback on diabetes care to general practitioners on the quality of care in patients with Type 2 diabetes were eligible for inclusion in the review. Feedback was defined as "any summary of clinical performance of health care over a specified period of time". Electronic feedback was similarly defined, but delivered to the end user via computer.

Most included studies assessed printed feedback; only 20% of studies assessed electronic feedback; 10% assessed both printed and electronic feedback. Included studies were published between 1994 and 2005 (most after 2000). Most studies assessed patient-specific feedback and 20% assessed aggregated feedback for individual general practitioners or general practitioner practices. In just over half of the studies feedback was the only intervention; in the other studies feedback was part of a larger multicomponent intervention. Regular feedback was compared to other interventions and combined feedback was compared with other support interventions.

Single studies compared combined feedback with outreach visits, feedback with a face-to-face evaluation with an endocrinologist and feedback with written patient feedback in the waiting room plus a computer reminder.

Most studies were conducted in USA; others were in Scandinavia, New Zealand and The Netherlands. Trial duration ranged from two months to six years (median duration 12 months). All studies used process measures that formed part of routine diabetes management (such as blood glucose and serum cholesterol). Other outcomes included composite measures such as compliance rates and compliance to diabetes guidelines. Data were collected via chart review, encounter forms filled in by participating general practitioners, databases or by combinations of these methods.

Studies were primarily assessed by one reviewer. Any uncertainties were discussed with the review team members.

Assessment of study quality

The authors reported problems concerning randomisation, sampling, blinding and dropout, but did not report that they used a formal assessment of study quality.

Data extraction

It appeared that a single reviewer extracted study data. Data were divided into two groups dependent on whether they were process measures or outcome measures; composite measures were also noted. The authors reported types of data assessed and whether significant positive measures in favour of the intervention group were recorded. Where reported, effect measures that included odds ratios (ORs), relative risks (RRs) and hazard ratios (HR), with 95% confidence intervals (CIs), were extracted along with p-values for any differences in effects between intervention and control groups.

Methods of synthesis

Studies were grouped according to type of intervention and whether they assessed process measures or outcome measures. Study findings were combined in a narrative with supporting tables of summary data.

Results of the review

Ten RCTs (23,855 patients and 1,527 general practitioners) were included in the review. Numbers of patients within studies ranged from 359 to 5,628 and the number of general practitioners ranged from 35 to 484. There was no allocation concealment in any of the studies. All studies were unblinded; one study reported that it attempted to blind participants, but blinding was not effective. The authors reported that two studies had a risk of spill-over effect due to the included general practitioners all working for the same practice.

Feedback improved care for patients. One study reported no statistically significant changes. Nine studies reported a total of 23 statistically significant positive changes; 51 variables in these nine papers were not statistically significant.

Studies that reported statistically significant differences between feedback and control groups and reported positive effects in favour of feedback were: three trials for foot examination; one trial for Hba1c measurements; two trials for eye examination; one for influenza vaccination; two for cholesterol measurement (including one trial of electronic feedback); three trials of blood glucose measurement; one of dietary advice; two of micro albumin measurement; three of blood pressure levels (including one trial of electronic feedback); two of Hba1c levels; one of cholesterol measures; one of odds of receiving recommended care (electronic feedback study); one of compliance rate; one of recommended lab assays; and one of patient satisfaction.

Authors' conclusions

Feedback appeared to be a promising tool for quality improvement in diabetes care. Further research was required, especially to assess electronic feedback.

CRD commentary

This review answered a clear research question using well-defined criteria. A number of databases were searched for relevant studies. Inclusion was limited to studies written in English, so relevant data may have been missed and this was acknowledged by the authors. There may also have been a risk of publication bias as no specific attempts were made to locate unpublished data and the authors stated in their methods that they included only published studies; however, they also reported that one unpublished study was assessed for inclusion. There may also have been a risk of reviewer error and bias as only one reviewer appeared to have assessed studies for inclusion and extracted study data; studies were only referred to other reviewers if inclusion was unclear. Some methodological weaknesses in the included studies were reported by the authors, but the lack of a formal assessment of each individual trial made it difficult to assess data reliability. Given the variability in outcomes, interventions and study populations within the studies, use of narrative synthesis methods appeared appropriate. But, this made the interpretation of data and the significance of positive effects difficult. Therefore, overall the reliability of the findings was unclear and the authors' conclusions should interpreted with caution given the potential for missing data, variation between studies and an unknown risk of bias.

Implications of the review for practice and research

Practice: The authors reported that feedback could be a useful tool in diabetes management.

Research: The authors stated that randomised controlled trials that assessed feedback on diabetes care to general practitioners were required, especially to assess the effects of electronic feedback.


The review was funded by Regional and National Public Foundations, Vissings Foundation, Danielsens Foundation and A.P. Møllers Foundation Promoting Medical Science.

Bibliographic details

Guldberg TL, Lauritzen T, Kristensen JK, Vedsted P. The effect of feedback to general practitioners on quality of care for people with type 2 diabetes: a systematic review of the literature. BMC Family Practice 2009; 10:30: 10.1186/1471-2296-10-30. [PMC free article: PMC2690581] [PubMed: 19419548]

Indexing Status

Subject indexing assigned by NLM


Diabetes Mellitus, Type 2 /therapy; Family Practice /methods /standards; Feedback, Psychological; Humans; Physicians, Family; Quality of Health Care; Treatment Outcome



Database entry date


Record Status

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.

CRD has determined that this article meets the DARE scientific quality criteria for a systematic review.

Copyright © 2014 University of York.

PMID: 19419548


PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...