• No results found

PP 19: Adults who commence pharmacotherapy should have their medication adjusted as required and response assessed regularly (approximately 6-12 weekly) until sufficient improvement has been achieved or maximum tolerated

3. Literature review

The systematic literature review was undertaken according to the process outlined in the NHMRC Standards and Procedures for Externally Developed Guidelines (2007) by an external group from the Centre for Allied Health Evidence (iCAHE), University of South Australia, led by Dr Susan Hillier and Professor Karen Grimmer-Somers.

Searches were conducted in relevant databases using an agreed search protocol which lists details of search terms, inclusion/exclusion criteria, and data extraction and appraisal methodology. Additional hand searching was conducted by the NSF project team in several key journals to identify any major trials or meta-analyses published after the systematic literature review.

3.1 Criteria for considering studies for the review

Search dates

The search dates were 2006 to June 2010 for the first five questions relating to assessment of CVD risk which updated the search conducted for the Clinical Guidelines for the Assessment of Absolute Cardiovascular Disease Risk (which used no limits on the date of publication). The search dates were 2002 to June 2010 for the remaining questions relating to management of absolute CVD risk. Hand searching was conducted between June 2010 and May 2011.

Types of studies

Existing guidelines, systematic reviews (Level 1 evidence, based on the NHMRC Levels of Evidence and Grades for Recommendations for Developers of Guidelines (2009), randomised controlled trials (Level II evidence) were considered for inclusion, crossing intervention and diagnostic domains. Where there was a scarcity of Level I or Level II evidence, it was planned to expand the review to consider lower levels of evidence. Studies were limited to English language only.

Types of participants

The review included research conducted in adults without pre-existing CVD or in those with and without CVD but where those without CVD were reported separately.

Types of outcomes

In principle, the primary outcome for each question was cardiovascular events (definition for CVD as for the Guidelines for the Assessment of Absolute Cardiovascular Disease Risk).

The secondary outcome of interest was AR reduction, followed by surrogate outcomes such as individual risk factor reduction as specified in the questions (e.g. BP control).

3.2 Search strategy for identification of studies

A broad search strategy using the following databases and sources was used to identify potential studies:

• Medline

• Embase

• Cinahl

• PsychINFO

• Cochrane Library, including CENTRAL Cochrane Controlled Trial Register (CCTR) and DARE for some topics.

In addition, the following websites were searched including Australian Centre for Clinical Effectiveness, National Institute for Health and Clinical Excellence, National Library for Health, Swedish Council on Technology Assessment in Healthcare, US Agency for Healthcare Research and Quality, and the US National Guidelines Clearing House. The EWG were sent interim search reports and asked to identify any additional studies.

Hand searching undertaken after the online database searching included the following journals: British Medical Journal, New England Journal of Medicine, LANCET, Circulation, Journal of the American Medical Association, Archives of Internal Medicine, Medical Journal of Australia and Diabetes Care.

The Cochrane library was also reviewed to incorporate new or updated reviews. Hand searching was undertaken to identify major meta-analyses or landmark trials to maximise the currency of the text. In one situation, literature identified after the comprehensive literature review period was deemed by the EWG to be sufficiently important to result in a change to the recommendations (i.e. BP targets for those with CKD). This decision took into consideration the quality of evidence (all high-quality meta-analyses), the need to provide clinicians with the most useful recommendation, alignment to draft international CKD guidelines, and the likely scenario that the current guidelines could be out of date before they were published.

In addition to the initial searches, economic literature was searched via EBSCOhost database (Econlit & CINAHL), Ovid database (EMBASE, Medline), BioMed central and Cochrane library database (Health Technology Assessment, NHS Economic Evaluation). A broad search strategy of Australian and international literature (developed countries including European, North American and Canadian) for the years 2002–2010 was used. The cut-off dates build on the SIGN guidelines used during the systematic review phase.

Search terms

Search terms were used for each group of clinical questions/topics. Search terms were based on those reported in the Supplementary Guidelines Material (SIGN) where the first series of strings are disease/population identifiers and the additional strings relate to the specific question, i.e. intervention (e.g. alcohol and euphemisms).

Search strategies used in other databases were adjusted for different databases, but were substantially the same.

Searches were combined with guidelines, systematic review, and trial filters as appropriate.

3.3 Study selection

One reviewer assessed the titles and available abstracts of all studies identified by the initial broad searches (based on population and intervention) and excluded any clearly irrelevant studies. Two reviewers then independently assessed papers identified as potentially eligible studies using the inclusion criteria and resolved disagreements on inclusion by consensus, with reference to a third reviewer if necessary. This second phase thus focused on selection of studies based on the outcomes, treatment comparisons and any population subgroups (e.g. diabetes, CKD) which may have different effects of an intervention.

Hand searching identified 44 potential new trials or meta-analyses of which 9 were included in the final guidelines.

During finalising of the guidelines two further meta-analyses on BP treatment in those with diabetes were identified and included.

Search terms used in the economic literature review were essentially the same for each database. A broad population identifier (CVD or cardiovascular disease OR coronary disease OR heart attack OR stroke) was used followed by the following terms: Exp “cost and cost analysis”; Costs.

ti/ab; Cost effective$.ti/ab; Cost benefit analys$.ti/ab; Exp health care costs/; (economic adj2 evaluat$).ti/ab; and finally primary prevention. Additional snowballing searches were undertaken. The total number of hits was 204 of which 28 were considered in more detail by one member of the project team. Reviewing staff at Deakin University scrutinised the 16 abstracts for omissions and 9 additional appropriate papers were retrieved and reviewed.

The following criteria were used to select economic studies:

• overseas evidence in developed countries of Europe, UK, North America, Canada

• AR of cardiovascular disease criteria

• primary prevention population included has no previous history of CVD

• BP-lowering diuretics, beta blockers, CCBs, ACE inhibitors

• cholesterol-lowering medications statins

• antiplatelets (aspirin)

• adults 35–84

• health outcome measured in Disability Adjusted Life Years (DALYs) or QALYs.