Systematic review of the effects of care provided with and without diagnostic clinical prediction rules

Sharon L Sanders, John Rathbone, Katy J L Bell, Paul P Glasziou, Jenny Doust

Research output: Contribution to journalArticleResearchpeer-review

27 Downloads (Pure)

Abstract

Background
Diagnostic clinical prediction rules (CPRs) are worthwhile if they improve patient outcomes or provide benefits such as reduced resource use, without harming patients. We conducted a systematic review to assess the effects of diagnostic CPRs on patient and process of care outcomes.

Methods
We searched electronic databases and a trial registry and performed citation and reference checks, for randomised trials comparing a diagnostic strategy with and without a CPR. Included studies were assessed for risk of bias and similar studies meta-analysed.

Results
Twenty-seven studies evaluating diagnostic CPRs for 14 conditions were included. A clinical management decision was the primary outcome in the majority of studies. Most studies were judged to be at high or uncertain risk of bias on ≥3 of 6 domains. Details of study interventions and implementation were infrequently reported.

For suspected Group A Streptococcus throat infection, diagnostic CPRs reduced symptoms (1 study) and antibiotic prescriptions (5 studies, RR 0.86, 95% CI 0.75 to 0.99). For suspected cardiac chest pain, diagnostic strategies incorporating a CPR improved early discharge rates (1 study), decreased objective cardiac testing (1 study) and decreased hospitalisations (1 study). For ankle injuries, Ottawa Ankle Rules reduced radiography when used with clinical examination (1 study) but had no effect on length of stay as a triage test (1 study). For suspected acute appendicitis, CPRs had no effect on rates of perforated appendix (1 study) or the number of non-therapeutic operations (5 studies, RR 0.68, 95% CI 0.43 to 1.08). For suspected pneumonia, CPRs reduced antibiotic prescribing without unfavourable outcomes (3 studies). For children with possible serious bacterial infection, diagnostic CPRs did not improve process of care outcomes (3 studies).

Conclusion
There are few randomised trials of diagnostic CPRs, and patient outcomes are infrequently reported. Diagnostic CPRs had a positive effect on process outcomes in some clinical conditions; however, many studies were at unclear or high risk of bias and the results may be context specific. Future studies should seek to detail how the CPR might alter the diagnostic pathway, report effects on both patient and process outcomes, and improve reporting of the study interventions and implementation.

Trial registration
The protocol for this review was not registered with PROSPERO, the international prospective register of systematic review protocols. The review was conceived and protocol prepared prior to the launch of PROSPERO in February 2011.
Original languageEnglish
JournalDiagnostic and Prognostic Research
Volume1
Issue number13
DOIs
Publication statusPublished - 26 Apr 2017

Fingerprint

Decision Support Techniques
Outcome Assessment (Health Care)
Anti-Bacterial Agents
Ankle Injuries
Triage
Appendix
Appendicitis
Clinical Protocols
Pharynx
Streptococcus
Chest Pain
Bacterial Infections
Ankle
Radiography
Prescriptions
Registries
Length of Stay
Pneumonia
Patient Care
Hospitalization

Cite this

@article{9614a1e120234c09a2cb25ab61d2d621,
title = "Systematic review of the effects of care provided with and without diagnostic clinical prediction rules",
abstract = "BackgroundDiagnostic clinical prediction rules (CPRs) are worthwhile if they improve patient outcomes or provide benefits such as reduced resource use, without harming patients. We conducted a systematic review to assess the effects of diagnostic CPRs on patient and process of care outcomes.MethodsWe searched electronic databases and a trial registry and performed citation and reference checks, for randomised trials comparing a diagnostic strategy with and without a CPR. Included studies were assessed for risk of bias and similar studies meta-analysed.ResultsTwenty-seven studies evaluating diagnostic CPRs for 14 conditions were included. A clinical management decision was the primary outcome in the majority of studies. Most studies were judged to be at high or uncertain risk of bias on ≥3 of 6 domains. Details of study interventions and implementation were infrequently reported.For suspected Group A Streptococcus throat infection, diagnostic CPRs reduced symptoms (1 study) and antibiotic prescriptions (5 studies, RR 0.86, 95{\%} CI 0.75 to 0.99). For suspected cardiac chest pain, diagnostic strategies incorporating a CPR improved early discharge rates (1 study), decreased objective cardiac testing (1 study) and decreased hospitalisations (1 study). For ankle injuries, Ottawa Ankle Rules reduced radiography when used with clinical examination (1 study) but had no effect on length of stay as a triage test (1 study). For suspected acute appendicitis, CPRs had no effect on rates of perforated appendix (1 study) or the number of non-therapeutic operations (5 studies, RR 0.68, 95{\%} CI 0.43 to 1.08). For suspected pneumonia, CPRs reduced antibiotic prescribing without unfavourable outcomes (3 studies). For children with possible serious bacterial infection, diagnostic CPRs did not improve process of care outcomes (3 studies).ConclusionThere are few randomised trials of diagnostic CPRs, and patient outcomes are infrequently reported. Diagnostic CPRs had a positive effect on process outcomes in some clinical conditions; however, many studies were at unclear or high risk of bias and the results may be context specific. Future studies should seek to detail how the CPR might alter the diagnostic pathway, report effects on both patient and process outcomes, and improve reporting of the study interventions and implementation.Trial registrationThe protocol for this review was not registered with PROSPERO, the international prospective register of systematic review protocols. The review was conceived and protocol prepared prior to the launch of PROSPERO in February 2011.",
author = "Sanders, {Sharon L} and John Rathbone and Bell, {Katy J L} and Glasziou, {Paul P} and Jenny Doust",
year = "2017",
month = "4",
day = "26",
doi = "10.1186/s41512-017-0013-2",
language = "English",
volume = "1",
journal = "Diagnostic and Prognostic Research",
issn = "2397-7523",
publisher = "Springer",
number = "13",

}

Systematic review of the effects of care provided with and without diagnostic clinical prediction rules. / Sanders, Sharon L; Rathbone, John; Bell, Katy J L; Glasziou, Paul P; Doust, Jenny.

In: Diagnostic and Prognostic Research, Vol. 1, No. 13, 26.04.2017.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Systematic review of the effects of care provided with and without diagnostic clinical prediction rules

AU - Sanders, Sharon L

AU - Rathbone, John

AU - Bell, Katy J L

AU - Glasziou, Paul P

AU - Doust, Jenny

PY - 2017/4/26

Y1 - 2017/4/26

N2 - BackgroundDiagnostic clinical prediction rules (CPRs) are worthwhile if they improve patient outcomes or provide benefits such as reduced resource use, without harming patients. We conducted a systematic review to assess the effects of diagnostic CPRs on patient and process of care outcomes.MethodsWe searched electronic databases and a trial registry and performed citation and reference checks, for randomised trials comparing a diagnostic strategy with and without a CPR. Included studies were assessed for risk of bias and similar studies meta-analysed.ResultsTwenty-seven studies evaluating diagnostic CPRs for 14 conditions were included. A clinical management decision was the primary outcome in the majority of studies. Most studies were judged to be at high or uncertain risk of bias on ≥3 of 6 domains. Details of study interventions and implementation were infrequently reported.For suspected Group A Streptococcus throat infection, diagnostic CPRs reduced symptoms (1 study) and antibiotic prescriptions (5 studies, RR 0.86, 95% CI 0.75 to 0.99). For suspected cardiac chest pain, diagnostic strategies incorporating a CPR improved early discharge rates (1 study), decreased objective cardiac testing (1 study) and decreased hospitalisations (1 study). For ankle injuries, Ottawa Ankle Rules reduced radiography when used with clinical examination (1 study) but had no effect on length of stay as a triage test (1 study). For suspected acute appendicitis, CPRs had no effect on rates of perforated appendix (1 study) or the number of non-therapeutic operations (5 studies, RR 0.68, 95% CI 0.43 to 1.08). For suspected pneumonia, CPRs reduced antibiotic prescribing without unfavourable outcomes (3 studies). For children with possible serious bacterial infection, diagnostic CPRs did not improve process of care outcomes (3 studies).ConclusionThere are few randomised trials of diagnostic CPRs, and patient outcomes are infrequently reported. Diagnostic CPRs had a positive effect on process outcomes in some clinical conditions; however, many studies were at unclear or high risk of bias and the results may be context specific. Future studies should seek to detail how the CPR might alter the diagnostic pathway, report effects on both patient and process outcomes, and improve reporting of the study interventions and implementation.Trial registrationThe protocol for this review was not registered with PROSPERO, the international prospective register of systematic review protocols. The review was conceived and protocol prepared prior to the launch of PROSPERO in February 2011.

AB - BackgroundDiagnostic clinical prediction rules (CPRs) are worthwhile if they improve patient outcomes or provide benefits such as reduced resource use, without harming patients. We conducted a systematic review to assess the effects of diagnostic CPRs on patient and process of care outcomes.MethodsWe searched electronic databases and a trial registry and performed citation and reference checks, for randomised trials comparing a diagnostic strategy with and without a CPR. Included studies were assessed for risk of bias and similar studies meta-analysed.ResultsTwenty-seven studies evaluating diagnostic CPRs for 14 conditions were included. A clinical management decision was the primary outcome in the majority of studies. Most studies were judged to be at high or uncertain risk of bias on ≥3 of 6 domains. Details of study interventions and implementation were infrequently reported.For suspected Group A Streptococcus throat infection, diagnostic CPRs reduced symptoms (1 study) and antibiotic prescriptions (5 studies, RR 0.86, 95% CI 0.75 to 0.99). For suspected cardiac chest pain, diagnostic strategies incorporating a CPR improved early discharge rates (1 study), decreased objective cardiac testing (1 study) and decreased hospitalisations (1 study). For ankle injuries, Ottawa Ankle Rules reduced radiography when used with clinical examination (1 study) but had no effect on length of stay as a triage test (1 study). For suspected acute appendicitis, CPRs had no effect on rates of perforated appendix (1 study) or the number of non-therapeutic operations (5 studies, RR 0.68, 95% CI 0.43 to 1.08). For suspected pneumonia, CPRs reduced antibiotic prescribing without unfavourable outcomes (3 studies). For children with possible serious bacterial infection, diagnostic CPRs did not improve process of care outcomes (3 studies).ConclusionThere are few randomised trials of diagnostic CPRs, and patient outcomes are infrequently reported. Diagnostic CPRs had a positive effect on process outcomes in some clinical conditions; however, many studies were at unclear or high risk of bias and the results may be context specific. Future studies should seek to detail how the CPR might alter the diagnostic pathway, report effects on both patient and process outcomes, and improve reporting of the study interventions and implementation.Trial registrationThe protocol for this review was not registered with PROSPERO, the international prospective register of systematic review protocols. The review was conceived and protocol prepared prior to the launch of PROSPERO in February 2011.

U2 - 10.1186/s41512-017-0013-2

DO - 10.1186/s41512-017-0013-2

M3 - Article

VL - 1

JO - Diagnostic and Prognostic Research

JF - Diagnostic and Prognostic Research

SN - 2397-7523

IS - 13

ER -