EVALUATION OF CLINICAL PRACTICE GUIDELINES IN CHIROPRACTIC CARE: A COMPARISON OF NORTH AMERICAN GUIDELINE REPORTS
 
   

Evaluation of Clinical Practice Guidelines
in Chiropractic Care: A Comparison of
North American Guideline Reports

This section is compiled by Frank M. Painter, D.C.
Send all comments or additions to:
   Frankp@chiro.org
 
   

FROM:   J Canadian Chiropractic Assoc 2001 (Sep); 45 (3): 141153 ~ FULL TEXT

  OPEN ACCESS   


Melissa Brouwers and Manya Charette


Clinical practice guidelines developed by the Canadian Chiropractic Association (CCA) and the Council on Chiropractic Practice (CCP) were evaluated by three independent appraisers using the most current version of the Appraisal of Guidelines for Research and Evaluation in Europe (AGREE) Instrument. Eighteen eligible chapters within the two documents (nine from each organization) were evaluated. In general, the CCA document was rated more favourably than the CCP document. The strengths of both documents include clarity of objectives and target users and complete descriptions of methods used to formulate recommendations. Areas of improvement for both documents include the need for more detail regarding the bodies of evidence under consideration for each section of the guideline. This paper presents the complete results of the evaluation, discusses the strengths of each guideline document, and makes suggestions for areas of improvement.



From the FULL TEXT Article:

Introduction

Clinical practice guidelines are systematically developed statements designed to assist practitioner and patient decisions about appropriate health care for specific clinical circumstances. [1] In the area of chiropractic practice, two key clinical practice guideline documents have been produced in North America:

Clinical Practice Guidelines for Chiropractic Practice in Canada
commissioned by the Canadian Chiropractic Association (CCA), [2]

and the American counterpart,

Vertebral Subluxation in Chiropractic Practice, Clinical Practice Guideline
by the Council on Chiropractic Practice (CCP). [3]

The aim of both guideline documents is to facilitate an evidence-based approach to chiropractic care that enables conscientious, explicit and judicious use of the current best evidence in making decisions about the care of the individual patient. [4]

In 2000, a proposal was submitted and accepted by the CCA to evaluate the CCA and CCP clinical practice guidelines. The relative strengths and weaknesses between the guideline documents and within each document were considered, as well as their adherence to evidence-based principles.

The documents were evaluated using the most current version of the Appraisal of Guidelines for Research and Evaluation in Europe (AGREE) Instrument.s Although there is no universally accepted, validated tool for measuring the quality of practice guidelines, the AGREE instrument has been widely studied and is currently undergoing international validation as part of the BIOMED-PL96- 3669 project. One of the appraisers (MB) is a member of this collaboration and special permission was granted by Francoise Cluzeau, project coordinator of BIOMEDPL96-3669, to use the most current version of the instrument (May 1999). The AGREE instrument is generic and can be applied to practice guidelines that cover topics such as prevention, diagnosis, treatment or intervention in any disease area.

The appraisers are methodology specialists who have no experience or training with chiropractic practice or the specific clinical and policy literature in this area. The scope of this evaluation focused exclusively on the chapters in each of the guideline documents that met the criteria for appraisal set out by the instrument, with the emphasis directed towards the methodological quality of the documents. The evaluation did not include the appraisal of clinical care and policy considerations.



Methods

      Appraisers

The research team mandated to review the clinical practice guideline documents developed by the CCA and CCP was composed of three independent appraisers. The appraisers have expertise in the development of clinical practice guidelines for cancer care and in the implementation of the evaluation tool used in this study. The appraisers are methodological experts with training and experience in systematic review, critical appraisal, and study design. One of the appraisers (MB), a doctorate-level behavioural scientist, coordinated the overall project.

This initiative was funded by the CCA. None of the appraisers have had any previous professional relationships with the CCA or the CCP. The CCA and the project coordinator agreed upon the scope of the evaluation before the initiative began. The research team had complete editorial independence from the CCA in the execution of the evaluation process, the final report submitted to the CCA, and the writing of this paper.

      Evaluation instrument

It was agreed by the CCA and the project coordinator that the documents would be evaluated using the most current version of the AGREE Instrument. [5] The AGREE Instrument capitalizes on the large literature identifying attributes that define high quality clinical practice guidelines (CPGs). [6-10]Its purpose is to provide a framework to assess the confidence that the potential biases (e.g., methodological biases [11-13]) in guideline development have been adequately addressed, that the recommendations that emerge from the CPG are reliable and valid, and that practical issues have been addressed.

It is composed of 24 items organized into eight dimensions. The scope and purpose dimension consists of four items and considers definitions and descriptions of guideline objectives, target users, clinical questions and patient population. The stakeholder involvement dimension is measured by four items that address the membership of the guideline development group, external review of the guidelines, pilot testing, and the extent of patient involvement in the development of the guidelines. The dimension, identification and use of the evidence, consists of two items and considers the methods used to search for the evidence base and the criteria for selecting the evidence.

The formulating recommendations dimension is measured by four items and is concerned with the formulation of recommendations, links between the evidence and recommendations, consideration of possible benefits and risks, and the impact on resources. The clarity dimension consists of four items and addresses the guideline structure, recommendations, and options for care. The application dimension is focused on organizational bafriers, attitude/ behaviour change issues and tools for application, and consists of three items. Monitoring, consisting of two items, focuses on the criteria for adherence to the recommendations and the updating process. The editorial independence dimension consists of one item. The items are answered using a combined ordinal-dichotomous scale composed of six response options that includes a 4-point likert scale (points strongly agree, agree, disagree, strongly disagree) and two additional options no information to answer and not applicable.

      Scope of the evaluation

The project coordinator reviewed each of the guideline documents prior to beginning the evaluation. Each chapter was categorized as either eligible for evaluation according to the AGREE criteria, ineligible for evaluation according to the AGREE criteria, or as supporting (see Table 1). [Please refer to Full Text] Eligible chapters were those that focused on specific clinical care issues (e.g., diagnosis, treatment). Where possible, each of the eligible chapters from one document was paired with an equivalent chapter from the second document (e.g., CCA Frequency and Duration of Care was paired with CCP Duration of Carefor Vertebral Subluxation). The AGREE instrument was completed for each eligible chapter. Chapters that focused on non-clinical issues or issues outside the scope of the AGREE instrument were categorized as ineligible, and were not considered in the evaluation. Methodological issues such as the quality of the systematic review and adherence to evidence-based principles are essential components to high quality CPGs and are reflected in the AGREE instrument. [6, 7, 9, 14] These issue were often represented in what we have termed supporting chapters in the CCA and CCP guideline documents. Thus, for some items in the AGREE instrument, a general score was derived using the supporting chapters as a foundation on which the appraisers' response options were based. In the absence of additional information within the eligible chapters in each guideline document, the general scores for these items were applied. The evaluation focused on the written documents supplied by the CCA and CCP. No efforts were made by the evaluators to verify the processes, participation and methods outlined in the guideline documents.

      Procedures

The research team initially met to review each of the items in the AGREE instrument, to clarify any ambiguities with item content, and to further operationally define items where, at face value, potential misunderstandings or inconsistencies in interpretation were possible. At the conclusion of that meeting, each appraiser was provided with copies of the CCA and CCP guideline documents, a list outlining the categories in which the chapters were placed (Table 1), and 18 copies of the AGREE instrument. The AGREE instrument was completed for the nine eligible chapters from each organization. For both the CCA and CCP documents, the appraisers first read the supporting chapters, followed by the first eligible chapter, and then completed the AGREE assessment tool. A consultation meeting followed this process to review the results of the first pair of assessments, to address inconsistencies in the application of questionnaire items, to further refine the operational definitions of problematic items, and to determine which of the instrument items could be classified as general.

Appraisers then completed a review and assessment of each eligible chapter pair as outlined in Table 1. [Please refer to Full Text] The order of evaluation within each pair was counterbalanced to avoid bias based on chapter order. Consultation meetings were held after the first, second and third chapter pairs.

      Treatment of the data

Scoring   To evaluate the strengths and weaknesses of the CCA and CCP documents, responses provided by each of the assessors for each of the guidelines were obtained. Strongly agree, agree, disagree and strongly disagree responses were scored 4, 3, 2, and 1, respectively. No information to answer responses were given a score of 0, with the assumption that each of the characteristics featured in the instrument contributes to the overall quality of the document. Thus, if there were no indication that the characteristic was featured in the process, this would reflect poorer quality. Further, a not applicable response was also given a score of 0. However there was only one occasion in which the appraisers used this latter option.

      Analyses

Data were analyzed using SPSS Version 10.0.5 for Windows.

Interrater reliability   The Kappa coefficient is a reasonable method for calculating the interrater reliability of categorical data, whereas the intraclass correlation coefficient (ICC) is an appropriate measure when continuous data are used. [15] One of the challenges of a combined ordinal/categorical scale is determining the most appropriate method for calculating interrater reliability. However, given the scoring rationale described above, the fact that four of the items included only the ordinal scale component, and that the not applicable option was chosen only once by the assessors, it was decided that an ICC would be the most appropriate and best estimate measure of interrater reliability.

Evaluation   Various evaluation scores were calculated. First, item scores across the appraisers were summed to give a total score for each dimension and the entire questionnaire. Second, these scores were compared to the maximum total score (mts). The mts was calculated by multiplying the number of appraisers by the number of items in the instrument component under consideration (e.g., dimension or complete questionnaire) by the highest possible score (i.e., strongly agree response with a score of 4). For example, the mts for the scope and purpose dimension is 48 (3 appraisers x 4 items in scope and purpose dimension x 4 highest possible score). Third, score means and standard deviations were calculated for each of the eight dimensions and the entire questionnaire for the eligible chapters across each of the appraisers. These values were calculated for the whole CCA and CCP guideline documents as well as each eligible chapter within these two documents.

Finally, rank order of the total scores of the guideline chapters is presented.



Results

      Interrater reliability and differences between raters' scores

Measures of interrater reliability across all items and chapters revealed an ICC of r = 0.76 (95% confidence intervals [CI] = 0.73-0.79) for the CCA guidelines and r = 0.77 (95% CI = 0.74-0.80) for the CCP guidelines. Thus, adequate reliability was achieved. [15]

A one-way analysis of variance was conducted to determine if there were statistically significant differences among the appraisers on mean evaluation scores across all items and all chapters in each practice guideline document. For the CCA document, a statistically significant main effect for appraisers was found, F(2,645) = 8.23, p < 0.001. The mean evaluation score was more positive for one of the appraisers (reviewer B: m = 2.78, sd = 1.09), than either of the others (Reviewer A: m = 2.31, sd = 1.34 and Reviewer C: m = 2.52, sd = 1.24). Although the means fell in a similar pattern, no statistically significant difference among appraisers on mean evaluation scores for the CCP guidelines was found, F(2,645), p = 0.12 (Reviewer A: m = 1.76, sd = 1.39; Reviewer B: m = 2.00, sd = 1.45; Reviewer C: m = 1.84, sd = 1.49). To increase reliability and address the biases in scoring, score means were used for the inferential statistics.

      Evaluation scores

Table 2

Tables 2 and 3 provide quantitative summaries of the evaluation results. Table 2 summarizes the distribution of responses for the CCA and CCP documents. Table 3 [Please refer to Full Text] includes the mean, standard deviation, and total scores as a function of instrument component (i.e., eight dimensions and the complete instrument) for each of the eligible chapters and the guideline documents as a whole. Also included in Table 3 are the maximum total score (mts) as a function of instrument component.

Distribution of scores   A chi-square analysis was conducted to determine if the distribution of responses reported in Table 2 was significantly different between the two groups (CCA v. CCP). A significant difference was found, X2 (4) = 138.44, p < 0.001. As can be seen in Table 2, both groups had a similar proportion of strongly agree responses (CCA = 18.5% v. CCP = 16.8%). In contrast, the evaluation re sulted in two and a half times as many agree responses for the CCA document relative to the CCP document (47.5% v. 19.4%), over 40% fewer disagree responses (14.4% v. 22.5%) and half as many strongly disagree and no information to answer responses (CCA 6.6% and 13.0% v. CCP 13.3% and 27.9%, respectively). Interestingly, the most common response option in the CCA evaluation was agree (47.5%) in contrast to the no information to answer response with the CCP evaluation (27.5%). For both groups, the strongly disagree response was the least common response option.

Table 4

Chapter score evaluation   Based on these findings, it is not surprising that the mean scores and total scores suggest a tendency for the eligible chapters in the CCA document to be rated more highly than the corresponding chapters in the CCP document. Although the ranges were the same, the minimum to maximum of total scores for eligible chapters in the CCA document were higher (171 to 201) and did not overlap with the CCP document scores (115 to 145). Table 4 summarizes the rankings of all eligible chapters based on the total scores of the instrument (including those with no chapter equivalent). There are two noteworthy features of these rankings. First, the total scores of the CCA chapters were consistently higher (Tables 3, 4). Second, there appears to be no correspondence in the rankings of the CCA and CCP chapter pairs. For example, the strongest eligible CCA chapter was Frequency and Duration of Care. Its CCP counterpart, Duration of Care for Vertebral Subluxation was one of the organization's weakest chapters (Table 4). In fact, the biggest difference in total scores and mean scores was found with this chapter pair. Similarly, the CCP counterpart of one of the weakest CCA chapters, Modes of Care and Management, was one of its strongest chapters, Modes ofAdjustive Care. Although the total score of the CCA version of this chapter was still higher than the CCP version, the relative ranked positions of these chapters were very different.

      Instrument dimension evaluation

General findings   For the overall CCA document, mean scores on the dimensions from strongest to weakest are:

scope and purpose (3.2),
clarity (3.0),
application (2.8),
formulating recommendations (2.6),
monitoring (2.6),
stakeholder involvement (2.1),
identification and use of evidence (1.8) and
editorial independence (0).

For the overall CCP document, the mean scores are:

scope and purpose (2.7)
clarity (2.7),
stakeholder involvement (2.5),
formulating recommendations (1.7),
monitoring (1.2),
identification and use ofevidence (1.0),
application (0.4), and
editorial independence (0).

It is interesting to note that the scope and purpose and clarity dimensions were the strongest dimensions for both the CCA and CCP. Similarly, identification and use of evidence and editorial independence were two of the weakest areas for both groups.



Discussion

The purpose of this project was to evaluate the CCA and CCP clinical practice guideline documents using the AGREE instrument. [5] The findings from the evaluation process indicate that the CCA guideline document was rated more favourably than the CCP guideline document. For both the CCA and CCP document, the least frequent response option was strongly disagree. However, for the CCA document the agree option was most frequent in contrast to the no information to answer option for the CCP document. Indeed, recall that we were unable to find information to answer over one quarter of the quality items related to the CCP document. The interval of total scores for eligible chapters in the CCA document was higher (171 to 201) and did not overlap with the interval of total scores in the CCP document (115 to 145). Table 4 presents the rank ordering of the eligible chapters in the CCA and CCP guideline documents. There appears to be little correspondence in the rank order of common-theme chapters between the two groups, suggesting that the subject matter of the guideline chapters did not consistently predict quality scores.

The CCA chapter scores were typically higher on the scope and purpose dimension than the CCP chapters, with the exception being the chapters that addressed modality of care and instrumentation (Table 3). [Please refer to Full Text] Typically, the differences between the organizations were modest, with the exception of the chapters that addressed safety and complications (i.e., CCA Contradictions and Complications and CCP Patient Safety), where there was a full point difference between the mean scores. The differences between the two groups were attributed primarily to the inclusion of additional information regarding clinical questions of interest and patient characteristics in the CCA chapter.

The CCP document scores were consistently higher than the CCA scores on the stakeholder involvement dimension due primarily to the greater involvement of patients and methods experts in the guideline development process (Table 3). The variation within the CCP document can be attributed primarily to the extent to which patient preferences were considered in each chapter. The CCA score was the same across all chapters because information used to answer the items in this section was found in the supporting literature.

The primary area for improvement for both the CCA and CCP guidelines falls within the identification and use ofevidence dimension. Both the CCA and the CCP scored quite low on this dimension. Neither group provided comprehensive descriptions of the methods used to search, select, and synthesize the evidence (Table 3).

The mean scores for formulating recommendations were higher for the CCA document compared to the CCP document, with the former providing more explicit and complete descriptions. At no time was the mean score of the CCP chapter higher than its CCA counterpart on this dimension.

The clarity dimension is one of the highest ranked dimensions for both the CCA and CCP documents. With the exception of chapters dealing with the duration of care, where the difference in means is greater than one point (CCA = 3.4 v. CCP = 2.1), the scores are very close between the groups (Table 3). [Please refer to Full Text]

The application dimension yields the greatest and most dramatic difference in scores between the CCA and CCP documents (Table 3). This can be attributed to the CCA document addressing some theoretical aspects of application and the inclusion of flow charts and algorithms for the user. There was no overlap between the groups on mean score ranges.

As with the application subdimension, substantial differences on scores between the two groups are found with the monitoring dimension (Table 3). The CCA document out performed the CCP document and there was no overlap between the mean ranges.

Finally, there was considerable debate among the appraisers regarding editorial independence. Financial support to develop the guidelines for the CCA document comes from professional associations and groups that would have an interest in the results of the project. Members of these organizations were involved in each step of the development process. The financial supporters of the CCP activities are unknown. In the absence of an explicit statement indicating editorial independence from the funding body, it was felt there was insufficient information to answer this question positively for either development group. The means and total scores for both the CCA and CCP document are zero for all of the chapters.

An important consideration is whether the differences in quality favouring the CCA document are meaningful and important. Three points bear on this issue. First, the greatest differences in total scores between the CCA and CCP documents are with the dimensions application (205 v. 29), formulating recommendations (258 v. 153), and monitoring (123 v. 55). Within these dimensions, the differences can be attributed, in large part, to the absence of information in the CCP guideline document rather than a description indicating that the guideline developers undertook a faulty or weak methodology. The distinction between poor guideline process and poor reporting standards is important. Based on the argument above, there is at least evidence to support the notion that the CCA reporting style is more complete than the CCP reporting style.

It is anticipated that this style of presentation, as it was used in the CCP document, was purposeful. In the introductory statement, the CCP document indicated a commitment to making a "user friendly" compendium. Indeed, the actual text of the document was short and the use of bold typeface successfully highlighted the recommendations. However, this streamlining approach by the CCP may have inadvertently neglected the inclusion of very important information that would enable the reader to make informed assessments about the recommendations. The "user friendly" compendium perhaps could have served as a supporting document to one in which all the issues were addressed.

The second aspect that bears on this issue is related to perceptions about the relevance of the individual dimensions. There may be some debate among the guideline development groups regarding the relevance of each dimension of the AGREE instrument within the clinical, policy, and health services contexts in the chiropractic community. For example, the greatest discrepancies between the CCA and CCP documents were with the application dimension and the monitoring dimension. If these dimensions are considered less relevant by the chiropractic community than other dimensions in which there is greater correspondence, it could be concluded that the differences in scores between the two reports might be less meaningful than believed at first glance. Many factors can influence this debate: the guideline model of the development groups, the mandate and responsibilities of development groups in contrast to the mandate and responsibilities of the professional organizations, the expertise of the members of the groups (e.g., clinical experts vs. methodology experts vs. implementation experts, etc.).

Finally, examining response patterns is the third consideration when trying to understand the evaluation outcome differences between the two groups. The CCA document received twice as many positive responses (combining strongly agree and agree options) and half as many negative responses (combining strongly disagree and disagree) during the evaluation than did the CCP document. The larger proportion of no information to answer response options in the CCP evaluation results can not completely account for this difference. Further, higher scores were found across most of the dimensions for the CCA document, not only those judged as less relevant by the appraisers. Thus, there is evidence that absolute qualitative differences also played a role. In summary, higher ratings of the CCA document relative to the CCP document can likely be explained by a combination of reporting style and guideline quality factors.

The current evaluation shows that there is correspondence between the two documents regarding common areas of strength and common areas of weakness. Considering areas of strength, Grilli and his colleagues suggest high quality CPGs from specialty societies should report relevant stakeholder groups, a strategy to identify primary evidence, and a process to grade recommendations based on the strength of studies incorporated into the report. [14] The CCP and CCA documents meet two of these criteria; in both cases, a very elaborate and well thought out system of classifying the recommendations and the type and quality of research used to inform them is outlined. Further, comprehensive memberships of the guideline development groups, particularly in the case of the CCP, and clearly identified target audiences, added credibility to the processes.

In both documents, the reader could typically follow the rationale of why topics were chosen and understand the objectives. The documents were well organized and framed and the recommendations were easily located. Thus scope and purpose and clarity dimensions were the stronger features of both guideline groups that were well represented.

The third quality component advocated by Grilli was not successfully incorporated into either of the guideline documents. [14] The appraisers agreed that both guideline documents could be improved by being more explicit, particularly in the identification and use of evidence and the linkages between evidence and recommendations. Indeed, the feature that most clearly deviated from an evidencebased perspective of guideline development, was the lack of information describing the processes used to identify and choose the reviewed literature. Indeed, both developers did not explicitly detail the strategy used to search for the evidence, and neither outlined the specific inclusion and exclusion criteria used to select the literature. Thus, for each applicable chapter, the specific body of literature considered, how this literature was chosen, and why other literature was ignored was unclear. Further, there was considerable inconsistency across the chapters, for both groups, with the link between evidence and the recommendations.

An implicit rather than explicit use of evidence not only compromises the judicious consideration of and likely compliance with CPG recommendations, it leads to difficulties in replicating the guideline development process and reconciling differences in recommendations produced on the same topic by the two developers. [7, 9] For example, the recommendations for the CCA Frequency and Duration of Care chapter are very different from the CCP counterpart, Duration of Care for Vertebral Subluxation.

Whereas the former provides very detailed timelines, the latter explicitly indicates that there are no data to substantiate specific time periods for care. Review of the reference lists indicates very little correspondence between the two groups. This difference can not be explained by the different guideline dates (i.e., the more recent CCP document, does not have a reference list that only incorporates literature published after the release of the CCA document). The absence of information describing the literature search and selection process makes it difficult to comment on which one of the two sets of guideline recommendations is more valid. Further, the factors that could explain why such very different recommendations emerged are unknown (e.g., regional differences in care, differences in patient values or clinical culture, etc.).

Finally, there were times when the reports did not establish a link between the evidence and the recommendations, did not indicate the range of clinical options available, and offered ambiguous recommendations that provided little direction for the management of care, features linked to recommendation and uptake. For example, in an observational study exploring factors that facilitate and hinder CPG practice, Grol and his colleagues found almost a twofold difference in uptake between CPGs with ambiguous recommendations versus clearly worded recommendations (36% vs. 67%, respectively). [7]

The level of editorial independence of the guideline development group from the funders was also unclear in both documents. Lack of editorial independence has the potential to significantly jeopardise the quality of the document. Consider the findings by Barnes and Bero who found a statistically and scientifically meaningful relationship between conclusions of review articles on the effects of passive ("second-hand") smoking and affiliations with authors. [13] Here, the investigators found that of the 37% of reviews that concluded no harmful effects of passive smoking, over 70% of these authors had affiliations with the tobacco industry. If the CCA and CCP groups were in fact independent from the funders, an explicit statement to this effect would alleviate the liability inherent in the ambiguity. If there is dependency with funds, this has the capacity to undermine the credibility of the documents.

In summary, the strongest areas for both documents include an identification of guideline objectives and clinical questions/themes, description of target users and patient populations, a clear guideline structure, credible members on the guideline development groups, and a complete description of the methods used to formulate the recommendations. Areas recommended for improvement in both documents include a more explicit and complete description of the specific bodies of literature under consideration, more detail regarding the systematic methods used to search the evidence, incorporating inclusion and exclusion criteria, statements describing the links between the evidence and recommendations, and greater consistency in creating specific unambiguous recommendations.

Finally, the membership of the research team charged with conducting the evaluation precluded the consideration of clinical content or context. As such, it may be useful to engage in the evaluation process again bringing this perspective to the table. Further, it might be useful to repeat this exercise when updated guideline documents developed by the CCA and CCP are released and when the final version of the AGREE evaluation instrument is available in the public domain.



References:

  1. Woolf SH.
    Practice guidelines: A new reality in medicine. I. Recent developments.
    Arch Intern Med 1990; 150:1811-1818.

  2. Henderson, D, Chapman-Smith, D, Mior, S, and Vernon, H.
    Clinical Guidelines for Chiropractic Practice in Canada
    Canadian Chiropractic Association, Toronto, ON; 1993

  3. Council on Chiropractic Practice.
    Clinical Practice Guideline: Vertebral Subluxation in Chiropractic Practice
    Arizona: Council on Chiropractic Practice, 1998.

  4. Hayward RSA, Wilson MC, Tunis SR, Bass EB, Guyatt G,
    for the Evidence-based Medicine Working Group.
    Users' guides to the medical literature. Viii.
    How to use clinical practice guidelines.
    A. Are the recommendations valid?
    JAMA 1995; 274:570-574.

  5. The AGREE Collaboration.
    Appraisal of Guidelines for Research and Evaluation in Europe (AGREE) Instrument.
    May 1999.

  6. Cluzeau FA, Littlejohns P, Grimshaw JM, Feder G, Moran SE.
    Development and application of a generic methodology to asses the quality of clinical guidelines.
    International Journal for Quality in Health Care 1999; 11:21-28.

  7. Grol R, Dalhuijzen J, Mokkink H, Thomas S, Veld C, Rutten G.
    Attributes of clinical guidelines that influence use of guidelines in general practice: observational study.
    BMJ 1998; 311:237-242.

  8. Lohr KN.
    The quality of practice guidelines and the quality of healthcare.
    In: Guidelines in health care. Report of a WHO conference. January 1997,
    Baden-Baden: Nomos Verlagsgesellschaft, 1998.

  9. Browman GP.
    Evidence-Based Cancer Care and Clinical Practice Guidelines.
    Am Soc Clin Oncol 1998; 451-457.

  10. Brouwers M, Johnston M, Browman G (in press).
    Results of a prospective study to keep guidelines current.
    Proceedings of the ISTAHC 2001 Conference.

  11. Moher D, Pham B, Jones A, Cook D, Jadad A, Moher M, Tugwell P, Klass T.
    Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses?
    Lancet 1998; 352:609-613.

  12. Colditz GA, Miller JN, Mosteller F.
    How study design effects outcomes in comparisons of therapy. I: Medical.
    Statistics in Medicine 1989; 8:441-454.

  13. Barnes DE, Bero LA.
    Why review articles on health effects of passive smoking reach different conclusions.
    JAMA 1998; 279:1566-1570.

  14. Grilli R, Magrini N, Penna A, Mura G, Liberati A.
    Practice guidelines developed by speciality societies: the need for critical appraisal.
    Lancet 2000; 355:103-106.

  15. Streiner DL, Norman GR.
    Health measurement scales. A practical guide to their development and use.
    Second edition. Oxford: Oxford University Press, 1995.



Return to the GUIDELINES Section

Since 1-12-2002

                  © 19952023 ~ The Chiropractic Resource Organization ~ All Rights Reserved