Answers to reviewer’s report.
Clare Liddy
Use of secondary prevention pharmacotherapy after first myocardial infarction in patients
with diabetes mellitus.
C.H. Jørgensen, G.H. Gislason, O. Ahlehoff, C. Andersson, C Torp-Pedersen, P.R.
Hansen.
Text in black: reviewer’s report
Text in green: answers from the authors
Text in blue: changes to the manuscript.
Major compulsory revisions:
Methods: It would be of value to have a specific sub-section describing the local setting.
This was done briefly in the study limitations section; however, it would be helpful to have
a section earlier on the manuscript that describes the healthcare system in Denmark, the
total number of hospitals, pharmacotherapy payment structures for patients (for example
are all drug costs covered by the government?), etc. This would help the reader who has
no knowledge of the Danish healthcare system.
Answer:
The setting of the health care system in Denmark is now explained earlier in the
manuscript:
Page 6, paragraph 1, line 1
Health care system in Denmark
All permanent residents in Denmark can use the Danish health care system freely and are
entitled to free treatment at a hospital. Data in the present study were obtained from
admissions to all of the 82 existing public hospitals in Denmark in 1996. Expenses for the
cost of drugs are partially reimbursed. The more expenses patients have for reimbursable
drugs, the more reimbursement they will receive. Annual medical expenses >900 Danish
Kroners (DKr) are reimbursed by 50%, >1,470 DKr by 75%, and >3,180 DKr by 85%.
Moreover, general practitioners can apply (usually successfully) for 100% reimbursement
of drug expences payable to their patients with chronic conditions, i.e. DM.
Statistical analysis: Did the authors consider the clustering of patients within hospitals?
Again, we do not have any information of how many hospitals (clusters) there might be.
This would change the analysis approach and may impact the results of this study.
Answer: The use of secondary prevention therapies after first MI according to the specific
hospitals that patients were admitted to in Denmark has been investigated previously and
was found to vary substantially. In the present study, data from admissions to a total of 85
hospitals were included and we agree that clustering of prescription claims, e.g., with
patients discharged from highly specialized cardiology clinics receiving more evidencebased drugs than those discharged from community hospitals, is likely to have influenced
the results. However, an investigation of the effects of the clustering phenomenon was not
the scope of our current work and will require complicated analyses of a massive load of
data. Moreover, these analyses will not change the overall message that on a national
scale, many patients do not appear to receive appropriate post-MI secondary prevention
drug treatment. We therefore kindly suggest that studies of effects of prescription
clustering are reserved for future work, but this important limitation to the interpretation of
our data must clearly be added to the paragraph on study limitations in the manuscript.
Accordingly, the following changes have been made:
Page 22, paragraph 1, line 5:
The use of secondary prevention therapies after first MI according to the specific hospitals
that patients were admitted to in Denmark has been investigated previously and was found
to vary substantially (46). In the present study, data from admissions to a total of 82
hospitals were included and clustering of prescription claims, e.g., with patients discharged
from highly specialized cardiology clinics receiving more evidence-based drugs than those
discharged from community hospitals, is likely to have influenced the results. Therefore,
our results apply to an average of Danish hospitals and patients and cannot be
extrapolated to represent the performance status of any specific institution.
New ref. no 46:
Rasmussen S, Abildstrom SZ, Rasmussen JN, Gislason GH, Schramm TK, Folke F, et al.
Hospital variation in use of secondary preventive medicine after discharge for first acute
myocardial infarction during 1995-2004. Medical care. 2008;46(1):70-7. Epub 2007/12/29
Statistical analysis: Is there pharmacotherapy coverage for all patients in Denmark? I was
curious as to why the authors decided not to adjust for socioeconomic status.
Answer: Yes, there is pharmacotherapy coverage for all patients in Denmark (see our
response to reviewer’s concern #1 above). In other countries and settings we agree that
socioeconomic status is of greater importance regarding medication adherence. Additional
analyses adjusting for household income at the year of admission for MI as a proxy for
socioeconomic status did not affect the estimates for any of the investigated therapies and
therefore we kindly suggest that the model used in the original manuscript is maintained.
Supplementary data on this issue can be provided at the request of the editor.
4- Discussion Section:
The discussion section was very long and included description of all results. The main
message and results (use of secondary prevention pharmacotherapy in pts with and
without diabetes) gets a bit lost. I think the manuscript would be greatly improved by
focussing the on the big picture and the main messages that are trying to be conveyed. I
would find it a stronger paper if there was also some thoughts about how the care might
be improved given the obvious gaps. I would suggest that the gender and age differences
may be better presented in a second paper and then the authors’ can more fully discuss
the meaning of these particular results and also how this compares to the current literature.
We agree that the discussion was extensive but it was important for us to address the
differences found for age and gender. Indeed, previous studies of medical adherence after
MI also included gender and age results. We therefore kindly suggest that the data are not
broken up into two papers.
The following changes have been made to the manuscript regarding ways to improve
adherence:
Page 21, paragraph 2, line 7
Adherence is a key factor associated with the effectiveness of all pharmacological
therapies not least for medications prescribed for chronic conditions like cardiovascular
diseases and DM. Medical adherence is a complex issue and involves demographic,
psychological, and social factors, as well as factors related to the health care provider,
medical system, specific disease and treatments. Improved adherence may be achieved
by targeted education of patients, involving patients in decisions regarding their treatments,
changes of dosing regimens, reduction of adverse effects, and improvements of access to
medication and medical consultations.
We also agree that it is important to compare the our data with results from the current
literature and have therefore made following changes to the manuscript:
Page 20, paragraph 2, line 11.
Numerous previous studies have examined the proportion of hospitalized cardiac patients
discharged with secondary prevention medications. A meta-analysis of data from 376,162
patients in 20 studies of primary and secondary prevention (9 studies) assessing
adherence to aspirin, ACEIs, ARBs, β-blockers, calcium channel blockers, thiazides, and
statins found a mean general adherence to secondary prevention therapies of 66% (44). A
registry study of a 2008-2009 population cohort in Spain of patients admitted with acute
coronary syndrome found the proportions of drug adherence to be 69.9%, 45.4%, 43.3%,
and 58.8%, for antiplatelet drugs, ACEIs/ARBs β-blockers and statins, respectively (45). In
that study only 47.6% received 3 or more drugs and no association was found between
low drug adherence and DM (60). Although we did not specifically examine drug
adherence but only used the correlate of claimed prescriptions, the use of ACEIs/ARBs, βblockers and statins in the late study period (2003-2006) in the present study is consistent
with these more recent results albeit that the percentage of patients that used
combinations of 3 or more drugs (75,1%) in our study was markedly higher.
Minor revisions:
2- Methods Section:
- In the “Database” section, there needs to be an ‘and’ between percutaneous coronary
interventions (PCI) and coronary by-pass grafting (CABG) procedures.
We apologize and have made the correction as requested.
- Define STEMI acronym
We apologize and have defined the STEMI acronym, i.e., ST segment elevation
myocardial infarction (STEMI)
- Results section: Error on page 9, line 8 – should read ‘3,951 instead of 3.951
We apologize and have made the correction.