Reimbursement Restriction and Moderate Decrease in Benzodiazepine Use in General Practice

Joëlle M. Hoebert; Patrick C. Souverein; Aukje K. Mantel-Teeuwisse; Hubert G. M. Leufkens; Liset van Dijk

doi:10.1370/afm.1319

Published eLetters

If you would like to comment on this article, click on Submit a Response to This article, below. We welcome your input.

Submit a Response to This Article

Compose eLetter

Title *

Author Information

Contributors
First Name and Middle Initial * First or given name, e.g. 'Peter'. Last Name * Your last, or family, name, e.g. 'MacMoody'. Email Address * Your email address, e.g. higgs-boson@gmail.com Occupation * Your role and/or occupation, e.g. 'Orthopedic Surgeon'. Affiliation * Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.

Statement of Competing Interests

Competing interests? *

Yes

Please describe the competing interests

CAPTCHA

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

What code is in the image? *

Enter the characters shown in the image.

Vertical Tabs

Jump to comment:

Author response Re:Lack of credible data for policy-making
Liset van Dijk

Published on: 20 March 2012
Lack of credible data for policy-making
Christine Y. Lu

Published on: 19 January 2012

Published on: (20 March 2012)
Page navigation anchor for Author response Re:Lack of credible data for policy-making
Author response Re:Lack of credible data for policy-making
Liset van Dijk, Utrecht, The Netherlands
Other Contributors:
In response to the comments made by Christine Y. Lu we would like to state the following:
In the last decennia, overall trends in benzodiazepine use have been slowly increasing over time, especially in the Dutch population aged 55-64 (1). SFK, a pharmacy database covering the whole Netherlands, showed a sharp overall decrease in the first six months of 2009, in the first half year the policy measure (exclusion o...
Show More

In response to the comments made by Christine Y. Lu we would like to state the following:
In the last decennia, overall trends in benzodiazepine use have been slowly increasing over time, especially in the Dutch population aged 55-64 (1). SFK, a pharmacy database covering the whole Netherlands, showed a sharp overall decrease in the first six months of 2009, in the first half year the policy measure (exclusion of coverage of benzodiazepines) was effective (1). This overall trend was confirmed in the general practice database we used in our study (results not yet published). After those six months there still is a decreasing trend, albeit being clearly less sharp. Given the increasing trend over the last decades it seems that the goal of the restriction, reducing and avoiding irregular (chronic) use, is fulfilled. In our study we decided to focus on incident cases of sleeping and anxiety problems. The reason for this is that according to Dutch general practitioners, quitting benzodiazepine use in chronic patients is a bigger challenge than reducing the number of patients starting with benzodiazepine use (2). With this in the back of our mind we have conducted the above mentioned study questioning ourselves whether the effect of the policy measure was on avoiding chronic use by decreasing the number of patients actually starting with benzodiazepines.
Touching upon the first comment raised by Dr. Lu, we do agree with her that interrupted time series are a useful tool when comparing longitudinal trends before and after a policy change and are good for sharply-defined interventions. However, another common analytical approach in longitudinal methods for pharmaceutical policy evaluation is survival analysis, the method we used in our study. This method is useful for studying longitudinal data on the occurrence of events (also known as "time to event" studies).
There are two basic options:
1) Pre-post analysis of the same population
2) Concurrent analysis of those subject to and not subject to a policy.
This design is feasible as no isolated comparison groups are likely to exist. An example of this approach conducted on the impact of a policy event is the study published by Stolk et al in 2008: 'Impact analysis of the discontinuation of reimbursement: the case of oral contraceptives'(3).
We do not fully agree with the second comment that our analysis did not account for possible anticipatory effects. If we would have been able to take into account only discontinuation based on the regulatory change (and not because of anticipatory responses), we might have found a higher difference in discontinuation rate between 2008 and 2009 (see discussion section). However, in another analysis (not yet published) we compared the prescription trends in the final 6 months of 2008, and the final 6 months of 2007 in order to see whether a "storage effect" occurred. This proved not the be the case. In addition, we only included new patients in our study. Therefore, the effect of the policy measure on the discontinuation rate of chronic patients is not of relevance.
Finally, our research aim was to investigate whether the number of patients initiating benzodiazepine use was affected by the policy measure. We do agree with Dr. Lu that the policy measure will have unintended outcomes. Therefore we did look at the unintended effect of switching to other treatment like the use of SSRIs. We do agree that it would be interesting to look at the use of other substitute drugs as well although we believe that switching to other drugs is not common as shown by recent data which show no increase in substitutable drugs such as quetiapine (Seroquel?).
We did not look at other unintended effects such as increases (or decreases) in use of health services, such as physicians or emergency room visits. This was because we only looked at the effect of the policy measure one year after the policy measure came into force and we are of the opinion that this period is too short to see an effect on the use of health services. Also, our primary care database did not include hospital admissions and with regard to referrals, part of the practices had incomplete data, reason why we decided not to include this.
1. Stichting Farmaceutische Kengetallen. Benzo's uit pakket; Pharmaceutisch Weekblad 20 juni 2008. (article in Dutch)
2. Van Geffen K., et al. Slechts 1 op 10 stopt na benzomaatregel. Pharmaceutisch Weekblad 2009. December 4. (article in Dutch)
3. Stolk P., et al. Impact analysis of the discontinuation of reimbursement: the case of oral contraceptives. Contraception; 2008 Volume 78, Issue 5 , Pages 399-404.
Competing interests: None declared
Show Less

Competing Interests: None declared.
Published on: (19 January 2012)
Page navigation anchor for Lack of credible data for policy-making
Lack of credible data for policy-making
Christine Y. Lu, Instructor
To the Editor: Hoebert and colleagues investigated the impact of a policy that excluded coverage of benzodiazepines (BZs) for anxiety and sleeping disorder.(1) However, this study has critical methodological flaws that may invalidate its findings.
Firstly, the conclusion that this reimbursement restriction has resulted in a moderately positive impact on BZs use lacks methodologic foundation. This study uses on...
Show More

To the Editor: Hoebert and colleagues investigated the impact of a policy that excluded coverage of benzodiazepines (BZs) for anxiety and sleeping disorder.(1) However, this study has critical methodological flaws that may invalidate its findings.
Firstly, the conclusion that this reimbursement restriction has resulted in a moderately positive impact on BZs use lacks methodologic foundation. This study uses one of the weakest research designs--pre-post with no comparison group design--that would not merit inclusion in a Cochrane Systematic Review of health care interventions or policies. This uncontrolled study cannot conclude that the observed small or no differences in the study outcomes between 2008 (pre-policy period) and 2009 (post-policy period) are causally related to the policy. The study could have easily used the interrupted time series analysis of longitudinal data; this strong quasi-experimental method controls for most threats to internal validity (e.g., secular changes in prescribing) because it adjusts for pre-existing trends in study outcomes that are unrelated to the policy.(2)
Secondly, the analysis did not account for possible anticipatory effects. The authors mention that the proposed reimbursement restriction was released in mid 2008; anticipatory responses by prescribers and patients before the policy came into force could bias the results toward the null or no effect of the policy. To address the possibility of anticipatory responses to the policy, the analysis could have treated the interval from mid 2008 to December 2008 (the month before policy implementation) as a separate phase from both the pre- and post-policy periods.
Thirdly, this study primarily considered medication use only. Also, while SSRI use was examined, these drugs represent only a small fraction of potential substitute medications for BZs for anxiety and sleeping disorder.(3) Many crucial intended and unintended outcomes of this policy remain unevaluated. Benzodiazepines are safe, effective treatments when used appropriately for anxiety, sleep, panic, and seizure disorders.(4, 5) To understand if the policy reduced inappropriate prescribing without denying legitimate medical use, determining how the policy affect potentially problematic and non-problematic BZ use, and use of potential substitute drugs (e.g., non-BZ sedatives such as zolpidem) would be valuable.(3) Furthermore, reductions in BZ use may also be followed by increases (or decreases) in use of health services, such as physician or emergency room visits, or hospital admissions; for instance, to treat serious discontinuation symptoms (e.g., recurrence of anxiety disorder, withdrawal reactions, seizures).(5) Understanding about shifts in health services may influence interpretation of changes in medication use and estimates of financial savings.
While I agree with the authors on the importance of evaluating the effects of coverage restrictions for pharmaceuticals, the weak data are not credible evidence for public policy or clinical decision-making.
References:
1. Hoebert JM, Souverein PC, Mantel-Teeuwisse AK, Leufkens HG, van Dijk L. Reimbursement restriction and moderate decrease in benzodiazepine use in general practice. Ann Fam Med. 2012;10(1):42-9.
2. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27(4):299-309.
3. Ross-Degnan D, Simoni-Wastila L, Brown JS, Gao X, Mah C, Cosler LE, et al. A controlled study of the effects of state surveillance on indicators of problematic and non-problematic benzodiazepine use in a Medicaid population. Int J Psychiatry Med. 2004;34(2):103-23.
4. Shader RI, Greenblatt DJ, Balter MB. Appropriate use and regulatory control of benzodiazepines. J Clin Pharmacol. 1991;31(9):781-4.
5. Bambauer KZ, Sabin JE, Soumerai SB. The exclusion of benzodiazepine coverage in medicare: simple steps for avoiding a public health crisis. Psychiatr Serv. 2005;56(9):1143-6.
Competing interests: None declared
Show Less

Competing Interests: None declared.