Reviews of "How best to fix a broken hip" by March et al.

The peer review of this article was conducted as a web-based discussion between peer reviewers selected by the editors of the MJA, the authors, the MJA editor for this paper and a small panel of consultants who were invited to observe and comment. The record of this discussion is now published with the article.

Readers are invited to comment further on the article or the review process.
Click here to read more about the eMJA Internet peer review study

DateComment
Fri Sep 25 17:00:59 1998
Index of entries

1. Review of the first submission
Welcome
Reviewer - Owen Williamson
Editor - Bronwyn Gaut
Reviewer - Flavia Cicuttine
Reviewer - Lu Mykyta
Reviewer - Albert Kirshen
Author - Lyn March
Editor - Bronwyn Gaut
Comment - Rob Weller
Author - Lyn March
Reviewer - Owen Williamson

2. Review of the revised article.

Reviewer - Flavia Cicuttine
Reviewer - Owen Williamson
For the Editors - Bronwyn Gaut - Decision

Welcome to one of the first review discussions of the MJA's second internet peer review study. Thank you to the authors, reviewers and consultant panel members who have agreed to participate in this experiment. We hope we are working towards better methods of peer review for all of us.

Let me apologise in advance for the technical glitches that are all too likely when attempting something like this for the first time. Please email me if you encounter any difficulties, or if you have any questions or comments about the process that you wish to make.


Craig Bingham, IPRS-2 study coordinator

Tue Oct 27 13:59:35 1998



How best to fix a broken hip - review by Owen Williamson

The main point:

The management of proximal femoral fractures could be improved by following evidence-based guidelines.

Scientific content and basis of manuscript:

Abstract

Par 10 states that guidelines for the management of hip fracture should be evidence-based to optimise functional outcome while minimizing hospital length of this stay. The authors do not provide data or references to support this statement.

Introduction

Par 12 states that by the year 2021 (the admission rate for hip fractures) is expected to more than double. The implication of the paper is that there will be a corresponding increase in the need for services and thus guidelines should be introduced to improve the efficiency with which these services are provided. I feel that the derivation of such an important statistic should be better described.

'What is the right thing to do' - Literature Review.

The design of the systematic review was very good, conforming to Cochrane Collaboration and NH&MRC guidelines.

There was insufficient detail in par 17 to enable replication of the search. I am aware however restrictions on publication space may preclude a more detailed description.
Was the search was limited to English language publications? When I performed a Medline search using the recommended search words and limited to randomised controlled trials published between 1966 and December 1995, 195 were listed. Time did not permit me to determine whether 98 of these were not relevant, leaving the 97 described in par 9 as meeting the inclusion criteria.

The authors cited 21 papers regarding thromboembolism prophylaxis. In a Cochrane review by Handoll et al (1), papers by 26 groups published between 1973 and 1996 were analysed. Whilst the authors of the current paper conclude that patients should receive LDH or LMWH, with a preference for the latter, Handoll et al found there was an increase in mortality associated with the use of particularly LMWH prophylaxis and recommended caution in the application of guidelines that recommend the use of injectable anticoagulants in patients with hip fractures. It is of some concern that these divergent recommendations may have arisen as the result of incomplete ascertainment of relevant studies.

Similarly, the authors have concluded that prophylactic antibiotics by vein should commence at induction of anaesthesia and continue for 24 hours. Controversy still exists regarding the need for more than one dose of IV antibiotics. They have not cited the several studies (3-8) that are still to be assessed as part of the Cochrane Collaboration Antibiotic prophylaxis: hip fracture protocol by Gillespie et al (2). Have they overlooked or excluded these studies?

The evidence and the treatment conclusions are described in detail in Table 1 and par 33-52. There is no detail on the potential clinical importance of each recommendation. It is stated in par 24 that odds ratios, 95% confidence limits and number needed to treat calculations were performed, but these results were not presented.

The authors conclude that prophylactic IV antibiotics and prophylactic anticoagulation are the only forms of care that improve outcome (Table 1). Other forms of treatment, with the exception of preoperative traction, surgical wound drainage and urinary catheterisation, appear promising, but require further evaluation. If the authors feel that these other forms of treatment should be included in evidence-based clinical pathways, then more specific evidence should be presented regarding likely efficacy.

The authors conclude that surgical wound drains are a form of care that have not been shown to have the effects expected from them, but which may require further attention (Table 1). Presumably these effects include reducing wound collections and wound infections. Varley et al (references 89,90) concluded that although drains do not late seroma formation following their removal, patients who received wound drainage showed better wound healing, as rated by the ASEPSIS scale and a reduced infection rate. It may be therefore, that any apparent lack of efficacy of drains in other studies is due to lack of power, rather than lack of effect. The authors' recommendation seems to be a reasonable compromise, pending further studies.

'Are we doing the right thing?' - Medical Record Audit.

'Are we doing the right thing' begs whether 'What is the right thing to do' defines best evidence-based practice.

Par 54 states that 729 consecutive admissions were audited and will be the subject of a more detailed report evaluating the implementation of the guidelines. Par 9 states that the audit was retrospective. Presumably therefore a different study population will be used to evaluate the implementation of the guidelines.

The methodology of the audit of practice in five acute Northern Sydney Area Health Service public hospitals during the 1993/94 financial year is not explicitly stated. Medical records were examined and data extracted, presumably initially by someone other than the authors and a subsequent independent audit performed by 'experienced reviewers'(par22). Were these the authors or medical records administrators, how was the validation performed and what were the intra- and inter-rater reliabilities?

Par 58 indicates that a complete data set was not available for each patient from each hospital. Was the dataset sufficiently complete for the summary statistics to be truly representative of the study population?

The data presented in par 54-66 provides an excellent summary with regard to the current practice of interventions outlined in Table 1 and pars 33-50, but may be better presented as a table.

Discussion

Par 70 appears to introduce anecdotal and unreferenced material regarding clinicians' responses to evidence-based practices. If clinicians' impressions were systematically recorded as part of this study, the instruments by which they were measured and an analysis of results should be documented. It might be a pedantic point, but a study that attempts to provide evidence-based guidelines to management, should discuss evidence not impressions.

Par 73 and par 74 describe and anaesthetic staff concerns about the combination of prophylactic anticoagulants and regional anaesthesia and comments on the polarisation among anaesthetists regarding regional anaesthesia and reduced mortality and morbidity. It is not clear whether these statements are based on anecdotal comments by the anaesthetic staff at the hospitals involved in the current study or represent the results of a systematic review of the appropriate literature. If the latter is the case, references should be cited.

It is of significance that although there is an extremely small risk if spinal haematoma, the US Food and Drug Administration has issued a Public Health Advisory calling attention to post-marketing reports of patients developing epidural or spinal haematomas with the concurrent use of LMWH and have asked manufacturers of LMWH to revise product information accordingly.

No reference is cited for the published meta-analysis mentioned in par 73. Presumably this refers to the study by Sorenson and Pace (reference 40).

No references are cited for the studies regarding optimal time of surgery described in par 74.

No references are cited for studies regarding optimal time for mobilisation described in par 75.

No references were cited to support comments regarding hospital administrators described in par 76.

No references were cited to support the statement that costs could be reduced by earlier transfer to rehabilitation from more expensive acute care ward. Whilst such transfers may reduce the cost to the acute care service, is there an associated reduction in costs to the patient or community?

No evidence apart from the personal communication cited in par 12 has been presented to support the proposition that there is an imminent epidemic of proximal femoral fractures. Spector TD et al (9) suggested that the increase in the number of hip fractures in England and Wales seen during the 1960 and 1970s was diminishing during the 1980s. This data is somewhat dated, but suggests the need for qualification of statements regarding continuing trends.

Originality of thought

Although systematic reviews and meta-analyses have been published regarding DVT prevention, surgical techniques and pre-operative traction for hip fractures (1, 10-14), to the best of my knowledge, this paper is original in so far as it attempts to collectively and systematically review the many interventions associated with the management of hip fractures in order to develop management guidelines.

Style, format overall presentation

The presentation is excellent, particularly with the summary provided by Table 1.

Thought should be given to constructing a table that outlines the results of calculations described in par 24, so as to give some guidance to the likely benefit achieved with the introduction of the recommended interventions.

The data presented in pars 54-65 may also be more clearly presented in a table.

Graphics (quality of illustrations, readability of figures and tables)

The tables presented are both necessary and very readable.

Reader interest

Although the subject of this paper is probably of most interest to orthopaedic surgeons and therefore suitable for a specialist publication, it presents a clear example of how to conduct a systematic review using Cochrane Collaboration and NH&MRC guidelines and follow a structured approach to health outcomes research as advocated by the NSW Health Department. As such, I feel the paper is of wider interest and sufficient importance to warrant consideration for publication in the Medical Journal of Australia.

Does this manuscript require further statistical analysis?

I feel the paper would be improved by the publication of the analysis described in par 24.

The summary statistics presented in pars 54-64 are sufficient.

Par 65 states that patients admitted on a Thursday were likely to spend an extra two days in the acute care facility. No statistical comparison is made between the lengths of stay quoted, but this is reasonable if it is felt that the difference is systematic in origin, for example rehabilitation facilities do not admit patients over the weekend, rather than chance alone.

Do you have any ethical concerns about this study?

No.

Final comments

Overall, I feel that this is an excellent paper and gives hope to those who are concerned about the lack of interest in an evidence-based approach to orthopaedic surgery. Subject to the minor gripes mentioned above, I feel that this paper is worthy of publication in the Medical Journal of Australia.

My only disappointment is that, despite the tremendous amount of work that has gone into this study, I still do not know whether the introduction of these guidelines will actually improve either patient or economic outcomes. Are the authors going to study the introduction of the guidelines by conducting a randomised trial comparing outcomes in patients treated by surgeon preference or by proposed guidelines? Is such a trial feasible, now that the surgeons in the studied hospitals are aware of the guidelines?

I look forward to future reports.

References

1. Handoll HHG, Farrar MJ, McBirnie J, Tytherleigh-Strong G, Awal KA, Milne AA, Gillespie WJ. Prophylaxis using heparin, low molecular weight heparin and physical methods against deep vein thrombosis and pulmonary embolism in hip fracture surgery (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

2. Gillespie WJ, Walenkamp G, Hoffman CW. Antibiotic prophylaxis in patients undergoing surgery for proximal femoral fracture (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

3. Burnett JW, Gustilo RB, Williams DN, Kind AC. Prophylactic antibiotics in hip fracture. A double-blind, prospective study. J Bone Joint Surg (Am) 1980; 62: 457-62.
4. Chiu KY, Ng KH, Fung B, Lau SK, Chow SP. A prospective randomized study on four regimes of antibiotic prophylaxis for hip fracture operations. (Abstract). J Bone Joint Surg (Br) 1993; 75B (Suppl III): 230-.
5. Jones RN, Reski WV. Single dose cephalosporin prophylaxis of 929 surgical procedures in a prepaid group practice: a prospective , randomized comparison of cefoperazone and cefotaxime. Diagn Microbiol Infect Dis 1987; 6: 323-334.

6. Nelson CL, Green TG, Porter RA, Warren RD. One day versus seven days of preventive antibiotic therapy in orthopaedic surgery. Clin Orth 1983; 176: 258-263.
7. Paul KJ, Hennig FF, Bartsch MM. Perioperative prophylaxis in orthopaedic and traumatic surgery with Ceftriaxone. Eur J Surg Res 1989; 21: 33-5.
8. Periti P, Jacchia E. Ceftriaxone as short-term antimicrobial chemoprophylaxis in orthopedic surgery: a 1-year multicenter follow-up. Eur J Surg Res 1989; 21: 25-32.

9. Spector TD, Cooper C, Fenton Lewis A. Trends in admission for hip fracture in England and Wales, 1968-85. BMJ 1990; 300:1173-1174

10. Parker MJ, Handoll HHG. Extracapsular femoral fractures: conservative versus operative treatment (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

11. Parker MJ, Handoll HHG. Extracapsular femoral fractures: conservative versus operative treatment (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

12. Parker MJ, Tripuraneni G. Extracapsular hip fractures: Surgical techniques for internal fixation (osteotomy, compression, reaming) (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

13. Parker MJ, Handoll HHG, Robinson CM. Gamma nail versus sliding hip screw for the treatment of extracapsular femoral fractures (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

14. Parker MJ, Handoll HHG. Pre-operative traction for fractures of the proximal femur (Cochrane Review). In: The Cochrane Library, Issue 3, 1998. Oxford: Update Software.

owen.williamsonATbigpond.com

Wed Oct 28 16:26:15 1998



Thank you Dr Williamson for your detailed and comprehensive review of this manuscript. While we wait for the other reviewers' comments, perhaps Dr Cameron (for the authors) could ask Professor Cumming to explain on what basis he estimates that the incidence of hip fractures will have doubled by the year 2021 (personal communication par 21).
Bronwyn Gaut, Senior Assistant Editor, MJA

Thu Oct 29 15:44:10 1998



Review by Flavia Cicuttini:
Thank you for the opportunity to comment on this article.

This article examines the evidence base for the very important clinical problem of management of fractured neck of femur.

This is an important piece of work that will significantly contribute to our knowledge in this area.

There are a few comments that I would like to make regarding the methodology. I understand the amount of work that has gone into this manuscript, but I have a number of points that the authors may wish to consider. Some could be addressed by a comment in the discussion.

Main comment:
This manuscript is likely to contribute significantly to the debate on clinical guidelines for this condition. The literature that is included is almost 3 years old. If this work is to contribute to guidelines, an update of the evidence, published in the interim, needs to be included. This could be added in an appendix.

Methods
Para 17:
120 articles were found and 23 were excluded. What were the inclusion critieria? What was the basis for the exclusions?
Only English language articles were used. Could this have limited the evidence available? Is it worth making a comment in discussion?

Para 19
Blinding of reviewers may be difficult as the format of high impact journals is recognized by most experienced reviewers. Did the reviewers have a 'check-list' or protocol to maximize objectivity?

Para 24
I may have missed it, but are the results of the meta-analyses for antibiotic prophylaxis and type of analesthesia included?

'Are we doing the right thing?'
Tabulation of these data may make the results easier to follow.

Para 54:
The authors have indicated that much of these data are the subject of future publication. Nevertheless we are provided with some data so need more information about the quality of the data provided. For example, does 729 subjects represent all admissions for hip fracture in that time period? Were there similar recruitment rates in the different hospitals? What are the results of the validation mentioned in para 22?


Discussion:
Is it worth making a comment about whether there is any evidence that clinical practice in some of the areas examined in this study may have changed since the survey was carried out 4-5 years ago?

I did not have any ethical concerns regarding the study.

Flavia Cicuttini

Fri Oct 30 14:03:50 1998



How best to fix a broken hip - review by Lu Mykyta

At the outset, let me say that I consider this paper well worthy of publication, and hope that it will engender a lively correspondence.

I realise that the paper literally deals with "fixing", rather than "managing", a broken hip. What is not dealt with is prevention of future fracture by attention to falls and osteoporosis. I am aware of the clinical work of my colleagues, and I am sure that these factors are not neglected. I suspect that the authors take it for granted that these considerations are part of geriatric consultation and the rehabilitation programs that they conduct. This is not a feature of the management of broken hips in all metropolitan hospitals across the nation, and some investigation of fall and fracture risk must be initiated during the orthopaedic admission.

As part of my homework for this review, I accessed the guideline "Management of elderly people with fractured hip" (Scottish Intercollegiate Guidelines Network, July 1997). This guideline was developed by a Guideline Development Group using the same methodology as the authors, and the first two recommendations are "Prevention of fractured hip by the prevention and treatment of osteoporosis" and "Prevention of fractured hip by prevention of falls."



Lu Mykyta

Wed Nov 4 23:49:00 1998
How best to fix a broken hip: - comments

OVERALL:
The article provides a timely, focussed, well-written structured review of the management of fractured hips.

SPECIFICS:
par 17 - The literature review was limited to English language and MEDLINE/CINAHL. In the discussion (par 68), the authors should comment on the limitations of this method. Particularly, the use of the non-English literature may add to the treatments reviewed and results obtained.

par 21 - Were patients with a SECOND hip repair or replacement excluded from this grouping in the audit? If not, their results should be represented separately.

par 25 - It is suggested the MJA publish the one page summaries on its web site for others to review, as the summaries related directly to the substance of the paper, and include reference to the web link in the print version of the paper.

par 53 - Only 1 table is published in the paper. The journal and the authors are encouraged to consider including the key elements from par 54 through par 65 in a table listing the interventions in Table 1, with a second column added to indicate the results of the audit.

General question:
How do these results compare with studies from other countries?

RECOMMENDATION:
Accept paper with revisions.
Albert J. Kirshen, MD, MSc, FRCPC

Thu Nov 5 13:11:09 1998
Thank you to all the reviewers for their time & efforts reading our article and their very detailed comments. It is always difficult when time marches on to know when to stop looking at the literature. This was first submitted 18 months ago.
My co-authors and I are meeting to respond to your comments and will post a reply by the end of next week. Thanks again
Lyn March
Lyn March

Thu Nov 5 16:04:48 1998
Bronwyn Gaut, for the Editorial Committee:

Thank you to the reviewers and our consultant panel. We would like to offer the authors the opportunity to revise their manuscript and will set a deadline for revision of 11 January 1999. Although we believe all the comments are relevant, some may not be able to be addressed within the MJA's publication limitations.

For example, Dr Williamson's suggestion of publishing the analysis described in par 24 may take up too much space. Another of his suggestions (presenting the data in par54-par66 in a table rather than as text) is excellent, but such a table might become unwieldy with footnotes and explanations - could the authors give it some thought?

In addition to addressing the reviewers' comments, we would like the authors to restructure the abstract to use the usual headings for reviews (ie, Objective; Data sources; Study selection; Data extraction; Conclusions).

The length of the article also needs to be borne in mind. We usually try to restrict articles to four pages in length (about 2500 words, plus figures and tables). That's going to be difficult in this case.

I would like to encourage further discussion and negotiation before the authors embark on their revision.
Bronwyn Gaut

Mon Nov 16 21:13:42 1998
I can only comment on the conclusions about rehabilitation, which appear reasonable. The statement that patients stay longer in acute centres if awaiting rehabilitation than if being discharged to Nursing Homes is against my experience.
Rehabilitation Assessment within 3 days of acute admission may be the ideal, but would be difficult to achieve with current resources and without education of orthopaedic services.
Rob Weller

Tue Jan 19 19:21:51 1999
January 19:
HAPPY NEW YEAR to all. Our apologies for being slow on the replies - various commitments and annual leave has seen us return to work only this week.

Reviewer one - Owen Williamson
1) we will move the statement re optimising outcome from conclusions but leave it in our discussion as a reasonable statement that doesn't require referencing.

2) the estimated hip fractures were calculated by Bob Cumming using ABS projected population figures and current age specific hip fracture rates and this will be explained and referenced in the text.

3) Search strategy will be outlined a little more within space limits - yes it was english language only, also age restricted .

4) The discrepancy with the Cochrane number of references was the inclusion to Dec 1996 - we don't believe we have a different conclusion however - the difference between a systematic review and a clinical guideline is that with a guideline you have to make a decision one way or the other - you cant say we cant decide how to treat you. Our description of the literature is the same.

5) re the antibiotic references. The additional Cochrane ones cited are known to us and mostly included abstracts, measures of antibiotic levels rather than the infection rates, and also a non-randomised study we exlcuded. We have produced another meta-analysis on this topic and have concluded that a single dose is probably all that is needed and will change this in the paper.

As Bronwyn Gaut has pointed out - it is not possible to provide all the references and all the individual odds ratios in this article.We have not done meta-analyses for all our summaries but we have provided individual paper reviews with results that would allow others to do them. These are published in our report that is referenced. We will provide this on the Northern Sydney Public Health Units web site and could quote the site in the article for people to access if they need further details - is this an acceptable alternative?
6)Owen requested info about the implementation of the guidelines - this is the subject of another paper - if we can get our guidelines published.

Following on from this is the request for more details about the audit process - as this is not central to this paper we will delete this.

7) The section relating to current practice can easily be included as an extra column to Table one. The Cochrane Treatment conclusion could be omited as it is difficult to interpret and we'll stick with the NH&MRC levels of evidence.
8) Discussion: about the anecdotal comments - yes we will state that this is what they were - I usually preface my talks about my experiences with evidence-based guidelines with a comment about the irony of having to resort to anecdote to convey the vastness of the task of developing and implementing such activities. We will state it more openly.

9) the heparin and spinal comments are very relevant - our original draft had very lengthy discussion on this - but given space limits it went - "first do no harm" but by the desire to avoid the one in a million haematoma - many others may die from their pulmonary embolus - we have acknowledged that it is controversial and will emphasise this more strongly.

10) we have all the references mentioned on par 73, 74 & 75 and can easily add them to reference list.

11) no reference for reduction of costs statement - we'll modify this to be a hypothesis - a very reasonable one we would have thought.

12) re par 24 comment - very difficult to interpret the data in this way as to likely benefit or value for money for following guidelines - we can state that we are not ready to say that yet but hope the access to the web site may help others deduce it for themselves in their own setting.

13) WE WILL ADD the pars 54-65 data as the extra column in table one as discussed.

14) Thank you for the very detailed reference list - these have been followed up.

REVIEWER 2 Flavia Cicuttini

1)Flavia has highlighted the problem of paper publication and having to go through this whole process -we first sent this to the MJA in June of 1997! We could list new RCTS in an appendix but we would not be able to put them through the same process that those included have been through - We believe the value of this paper is to give some insight to the process involved if a full clinical guidline for a whole patient admission rather than just one small part in isolation is to be developed. Very difficult to convey in this format and it may become more apparent in subsequent papers that describe the process and outcomes of implementation in more detail. Would the journal be interested in receiving the series??!!

2) we will make it more explicit as to inclusion criteria and will make a brief comment in discussion about potential limitations.

3) Tried to give the reviewers articles that were not their favourite area of interest. All followed a standard protocol, all were double read plus a consensus conference for ANY discepancies on the coding form - this is far more detailed than even Cochrane do for their reviews.

4) A summary of the antibiotic meta-analysis are available in the report which we can make available on the web site - full details of the process are the subject of another paper as we felt it far to detailed for the MJA and this article.

5) re discussion and change in clinical practice - this too is more relevant for our paper on implementation and practice change so will appear in their a subsequent audits have beendone in hospitals with and without the guidelines in place.

Reviewer 3 - Lu MYKYTA
Lu raises some very important issues that are beyond the scope of this article that already requires considerable editing to get into the word limit. - although our wider review addressed prevention of hip fracture , this article only refers to the acute clinical management as this is relevant for clinicians in the acute hospital setting.

Comment 4 Rob Weller
Thank you for comment - Ian C will discuss with you - we agree people get back to nursing homes faster - the comment about longer stay was for those who were waiting to be discharged to an off-site REHAB facility rather than those going to a rehab unit or ward that was in the same grounds as the acute hospital.

Reply to Bronwyn Gaut:
We will submit the changes as suggested, including the revised abstract in MJA format and will cut the whole article back as much as possible by deleting a lot of the discussion about deviation from current practice - although we believe this help justify to the clinician reader that guidelines may be useful.
The revised manuscript will be updated electronically by Wed Jan 27 - We hope this will still be acceptable to you.
Thank you for your interest in our article.

Lyn MARCh & co

Tue Jan 19 19:35:34 1999
Our apologies for not replying to Reviewer 4 - Albert KIRSHEN

par 17 - we will briefly discuss limitations of limiting searching - but the reader will be able to read the methods and have their own reservations or otherwise.

par 21 - we have combined them and would not be easily able to present them separately for this paper - we do know that about 10% have had a previous hip fracture.

par 25 has been discussed and we are open to suggestion of whether the MJA has summaries on web site or we put them on our local one.

par 53 yes thank you we will add this to discussion

re the other country comparison - this is a bit beyond this paper as well but we can briefly discuss a Uk audit that shows we are doing well in comparison but we still have room for improvement.


Lyn MARCH & CO

Tue Feb 2 17:01:25 1999
Thanks to Lyn March for her detailed response to my comments.

I am happy that the authors have satisfactorily addressed my concerns and look forward to seeing the amended paper in print.
Owen Williamson

Wed Feb 10 17:38:47 1999
I was intrigued, upon reflection, by Lyn March's comment that "the difference between a systematic review and a clinical guideline is that with a guideline you have to make a decision one way or the other - you cant say we cant decide how to treat you".

This statement seems to suggests that it is acceptable to ignore a lack of evidence in order to come to a clinical decision.

I agree that it is not helpful to tell patients we can't decide how to treat them, but it is reasonable to admit to them that we don't know the best way to treat them and to then discuss the pros and cons of alternative "reasonable" treatments.

As far as guidelines are concerned, if a decision to treat cannot be recommended on the basis of evidence, then this should be explicitly stated.

Similarly, if no evidence is available to support a particular treatment, then the rationale for recommending such a treatment in clinical guidelines should also be explicitly stated.

The authors have addressed this difficulty somewhat by defining the level of evidence used to support treatments recommended in their guidelines and at present I can't think of a better approach.

Thank you once again for your challenging comment.


Owen Williamson

Tue Feb 16 15:08:43 1999
Thank you Owen - I didn't express it very well but Yes that is exactly what the levels of evidence are intended to convey - individual clinicians may not follow level 3 & 4 evidence guidelines but wold have to have a good individual patient indication for not following level 1. As our guidelines were feeding into a pathway a recommendation for each step needed to be made.
Regards
LYN MARCH
LYN MARCH

Mon Mar 15 10:19:50 1999
The manuscript now addresses the previous concerns.
The only additional comment that I would like to make is that the
authors have not indicated how they think restrictng the information
to English publications and post 1996 work may affect the results.
This has been alluded to in para 78, but no indication of the
possible consequence of this has been made.
Otherwise, I think that the manuscript is excellent.


Flavia Cicuttini

Mon Mar 15 13:32:51 1999
The rewritten paper appears to address my previous concerns.

I would however like to make a few comments.

I was confused by par 56 which suggests that considerable variation in the process of care is tabulated in Table 1. Is the variation tabulated in another table that has been omitted or does Table 1 represent a process of care even if it contains practices that have been recommended against?

Although the paper alludes to the Report from the Northern Sydney Area Fractured Neck of Femur Health Outcomes Project obliquely through an MJA URL, no direct reference is made for those who do not have Internet access yet might wish to read the report. Please include a direct reference to the report in par 24 (or do the editors feel this would represent some unreasonable form of advertising?!).

I would still have some quibbles about the duration of prophylactic antibiotic use and whether or not the risks of prophylactic anticoagulation are outweighed by a reduction in mortality in the population concerned, but accept that, given the current state of knowledge, the guidelines appear reasonable and safe.

I would certainly support the publication of this paper subject to the above minor amendments.
Owen Williamson

Fri Mar 19 15:42:22 1999
MJA Editors' response to the authors

Thank you for your revision, which appears to have addressed the majority of the reviewers concerns, within the word limit constraints.

1. I would like to comment on some of Dr Cicuttini's original criticisms
a) The age of the data. The age of the data was certainly a concern to the editors - it is our usual policy to discourage work that is more than three years old, which certainly applies to the audit data (1993/94), and makes the literature review borderline. However, we believed that the work was important, that no other similar work had been done in Australia, and that the audit data, although quite possibly being different from current practice, was necessary as a baseline if subsequent audits looking for change in practice were to be done. The Jounal would certainly be interested in seeing further timely updates of the evidence.

Although Dr March et al replied that they had first sent this article to the MJA in June 1997, that's not quite how we see it. The article submitted then was rejected in October 1997; the current article is a rather different resubmission, received by the Journal in July 1998.

b) Inclusion and exclusion criteria - inclusion criteria have now been specified, but actual numbers of articles found, and any exclusions, have not. We would like the authors to add this.

c) Validation of the audit data - we don't think it is appropriate to just delete any mention of this - it is important to the methods and needs a brief explanation for those who can't easily access the full description in the report to the Area Health Service.

2. Dr Williamson's original comments
Most of these appear to have been addressed, except those he notes in his subsequent comments of 15 March. We note his comment about the "epidemic" of hip fractures (at the end of the discussion section) has not been addressed. We think it should be, or a less strong word chosen.

3. Dr Mykyta's original comments
A very important point. We accept Dr March et al's reticence to further expand their current study, but hope they might address this in future studies.

4. Dr Weller's original comments
We suggest that the word "off-site" be added to the description of rehabilitation facility to clarify this for others who may have the same concerns.

5. Dr Albert Kirshen's original comments
Comments appear to have been addressed within reason. As for the extra summaries, we are certainly able to post extra information about articles on our website. Dr March has supplied an electronic version of the detailed report and we are providing a link to this document.

Other Editorial Committee comments on revised manuscript:
These are relatively minor and, in our usual process, would probably have been left to the copyediting stage:
(i) Abstract - It is not our format to include an introduction in the abstract
- Results - please add the dates of the "review of current practice"
(ii) Introduction - When we quote ABS data, we normally give catalogue numbers of the source of the information - please add
(iii) Methods - "Are we doing the right thing?" - Please give the category names of ICD-9 codes and some descriptive terms to the procedural codes (for readers who are not familiar with them)
Development of evidence-based guidelines - first sentence - "the key steps . . . had been identified" - could the authors clarify this? Do you mean identified from the literature review?
(iv) Results - could the authors add in the reference numbers for each point (we know they're in the table, but it is also important that they are clear in the text)
- "Are we doing the right thing" - "antibiotics were used in the vast majority of all hospitals" - there were only 5 hospitals - should "vast majority of" be deleted?
We are prepared to accept the manuscript now, on the understanding that the remaining queries are dealt with promptly (which should be simple). If the authors are agreeable, we will open the peer review process to public comment now, with print publication to follow (after copyediting) in the next available issue. Further discussion is encouraged and changes to the manuscript can continue to be made until the article "goes to press".

Bronwyn Gaut, Senior Assistant Editor