SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References

Drug switching is a common medical practice. It indicates continuation of treatment regardless of the reason why the original therapy was stopped and switched. Therefore, the aims of this study were to develop a novel method for determining drug switching from routinely acquired NHS health data and to explore the aspect of continuation of care for patients. Patients who were first prescribed ramipril, simvastatin and an angiotensin receptor blocker (ARB) between 1 March 2004 and 28 February 2007 and discontinued their medication within 6 months of the index prescription were identified from the PTI database. The identified patients were then categorized into three groups: i) patients who were switched to a different drug for the same medical condition, ii) patients who were being prescribed with other types of antihypertensive/lipid-regulating drug prior to the initiation of study; and iii) patients who were without any continuation of care or therapy. Twenty percent (808), 29%(1429) and 14%(455) of the identified patients discontinued ramipril, simvastatin and ARB, respectively, within 6 months of an index prescription. Among the ramipril discontinuation group, 36.4% of the patients were switched to another antihypertensive, while another 31.6% of them were without continuation of care. In patients discontinuing ARB, 30.6% were switched, while another 30.1% were without continuation of treatment. In patients discontinuing simvastatin, 28.8% were switched to another lipid-regulating medicine, while another 63.1% of them were without continuation of care. The results of this study confirm that primary care prescribing databases can be used to determine drug-switching information and continuation of care/therapy.

The term ‘drug switching’ is used to denote a switch of drug therapy for the treatment of a particular medical condition to a different drug, which may be similar or belong to the same drug group or have the same therapeutic effect [1-4]. Reasons for drug switching are varied. Walker et al. [5] reported that adverse drug reactions (ADRs) and lack of efficacy were the main reasons for switching of non-steroidal anti-inflammatory drugs (NSAIDs). Other postulated causes include the ease and convenience of administering a particular drug formulation, cost to the patient, drug availability, tolerability of side effects, brand or drug preferences of patient/doctor and marketing strategy of individual drug companies [1-8]. Drug switching may also be undertaken as a national cost-saving exercise. A current example is the statin therapy. In the UK, statins represent a major cost to the National Health Service (NHS), with atorvastatin and simvastatin accounting for approximately 85% of all statin prescriptions [6]. Simvastatin, however, is now generic, and there is a significant cost-saving in the order of £2 billion over a 5-year period associated with the switch from atorvastatin to simvastatin in England alone [6].

Drug switching commonly involves the substitution of one drug with another from within the same drug class. It is a common notion among prescribers that if a drug is efficacious and tolerable, the same can be expected when a switch is made to a different drug in the same class. This, however, is not always the case as detailed by Furberg et al. [7] who identified the following three potential disadvantages of this practice:

  1. All drugs have multiple mechanisms or action. Focusing on a single beneficial mechanism may mask other mechanisms which may be harmful.
  2. It is inappropriate to assume therapeutic equivalence between drugs in the same class. An example would be the statins where fluvastatin has not been demonstrated to decrease mortality and morbidity owing to cardiovascular disease, unlike lovastatin, pravastatin and simvastatin.
  3. It is inappropriate to assume similar safety profile. The different safety profile of individual drugs within the same drug class is a well-proven issue.

Drug switching may also reduce patient compliance. Thiebaud et al. [8] reported that switching between statins may reduce patient compliance with statin therapy by up to 48% among patients.

A key aspect of assessing drug switching, which is frequently ignored, is the assessment of appropriate treatment continuation. When drug switching occurs, it may indicate a continuation of treatment regardless of the reason(s) for discontinuing the original therapy. On the other hand, a discontinuation of chronic prescription (e.g. antihypertensive agents) without drug switching indicates a complete discontinuation of therapy, which may result in an undesirable patient outcome. Treatment continuation has also been reported to improve patient outcome, especially in chronic diseases such as hypertension and hyperlipidaemia [9, 10]. If a patient diagnosed with hypertension is discontinued from antihypertensive therapy and is not switched to any other treatment, it indicates an absence of continuation of care for hypertension that may result in an undesirable patient outcome.

Continuing our previous research using routinely acquired National Health Service health data (practice team information database or PTI database) for pharmacovigilance (occurrence of adverse drug reactions/events) [11, 12], the aims of this study were to develop a novel method for determining drug switching from the PTI database and to explore the aspect of continuation of care for patients who were started on long-term therapy.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References

Determining drug switching from the PTI database

Using primary care data from the Practice Team Information (PTI) database, patients who discontinued from ramipril, simvastatin and angiotensin receptor blocker (ARB) prescription within 6 months following the index prescription were assessed. All seven types of ARBs were analysed as a single group. Patients who discontinued within the first month (1–31 days) were classified as early discontinuers, those who discontinued between the second and third month (32–95 days) were classified as intermediate discontinuers and those who discontinued between the fourth and sixth month (96–185 days) were classified as late discontinuers. The PTI database is an electronic database that contains routinely collected primary care healthcare data. Approximately 60 general practices contributed data to the PTI database, which included prescribing information for approximately 224,000 patients (6% of the Scottish population). Each patient was registered with a primary care physician. Practices contributing to the PTI databases were carefully selected to provide a reasonable reflection of the Scottish population with regard to age, gender, deprivation and urban–rural ratio mix. Information collected and stored within the PTI database included patient demographic data (non-personally identifiable), anthropometry data, diagnosis, prescriptions, practice visit details and referral details [13]. These medicines were chosen because they have well-established safety profiles and high frequency of use. Information regarding prescribing as well as patients' contacts with their primary care physicians within 5 years, from 1 September 2001 to 31 August 2006, was extracted. To determine accurate dates of the start and discontinuation of the prescriptions, the periods between 1 September 2001 and 31 August 2002 as well as between 1 September 2005 and 31 August 2006 were used as washout periods.

Patients started on the medicines of interest between 1 September 2002 and 31 August 2005 were identified and then tracked as an exposed cohort. Any records for apparently inactive patients (e.g. dead or moved from practice) were excluded from the analysis. Only active patients (i.e. patients who visited their practice for consultation/prescription prior to switching or after discontinuation of index prescription) were included. Following identification of these patients, the exact date of the last prescription for each study drug was determined, and the switched drug, if any, was identified.

Drug switching and switching period

Drug switching was defined as starting a different drug for the same medical condition to replace the discontinued drug of interest within a pre-defined period. The period between the last prescription of the drug of interest and the starting point of a different drug replacement is termed the ‘switching period’. For research purposes, the switching period should neither be too long that it might include drugs started for reasons other than switching, nor too short that appropriate medicines are excluded. Published studies frequently use switching periods between 60 and 120 days [1-3, 5]. Using the shorter 60-day switching period reduces the probability of detecting a drug started for reasons other than switching, but may exclude a significant number of switched drugs (i.e. high specificity but low sensitivity), and vice versa for the longer 120-day switching period.

A 60-day switching period was used in this study for early discontinuers (1-month discontinuers), and a 120-day switching period was used for the remaining two categories of patients (2- to 3-month discontinuers and 4- to 6-month discontinuers). This was done taking into account the unlikely scenario that a patient started on a new drug might be given a prescription for a period of more than 1 month without any subsequent follow-up within a short time frame for monitoring drug tolerability and suitability. Therefore, it would be safe to assume that any discontinuation of the drugs of interest within 31 days would be detected by the consulting general practitioners and any remedial action of switching the drug of interest would be taken within a short time frame (60 days from prescription date). On the other hand, the longer 120-day switching period was used for the other two categories as these patients had already completed the first month of treatment and might be issued a prescription for a longer period of 2 months and therefore necessitating a longer switching period of 120 days from the last prescription.

Continuation of therapy

Continuation of care or therapy was determined by identifying patients who were being prescribed at least one antihypertensive or lipid-regulating drug for the treatment of the same condition after a study drug was stopped (i.e. drug switch occurred). Patients without a drug switch but already on one antihypertensive or lipid-regulating drug were identified and categorized as patients with continuation of care/therapy. Only patients who discontinued a study drug without drug switch and were not on any treatment after being discontinued from a study drug were categorized as patients with complete stoppage of therapy.

Study approval

The project was reviewed and approved by the Primary Care Clinical Informatics Unit (PCCIU) advisory group [14] and did not require approval by the North of Scotland National Health Service Ethics Committees.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References

Data extraction from the PTI electronic prescribing database identified 4243 [54.7% men, mean age 64.4 years (S.D.: 13.3 years)] patients who were initiated on ramipril, 8849 [50.8% men, mean age 64.8 years (S.D.: 12.3 years)] patients who were on simvastatin and 3242 [43.5% men, mean age 64.5 years (S.D.: 12.8 years)] patients who were on ARBs during the study period between 1 September 2002 and 31 August 2005.

Ramipril

Approximately one-fifth (808) of the patients would discontinue ramipril within 6 months of the index prescriptions with 10% (421) of them discontinuing ramipril within 1 month, a further 5% (209) between two and 3 months and a further 4% (178) between four and 6 months throughout the study period (Table 1). 36.4% (294) of patients who discontinued ramipril within 6 months were switched to another antihypertensive, whereas 32.1% (259) of patients were already being prescribed with another antihypertensive prior to ramipril discontinuation. ARB was the most common switch drug, with 18.1% of patients switched to ARB once ramipril was discontinued. Approximately one-third or 31.6% (255) of patients were without continuation of care for hypertension, once ramipril was discontinued (Table 4).

Table 1. Ramipril drug-switching information according to antihypertensive drug class
Days on ramipril31 days or less (421 patients) (%)32–95 days (209 patients) (%)96–185 days (178 patients) (%)185 days or less (808 patients) (%)
  1. a

    9 were patients switched to two antihypertensives at the same time.

  2. b

    24 were patients switched to two antihypertensives and one patient were switched to four different antihypertensives.

  3. c

    5 were patients switched to two antihypertensives at the same time.

  4. Patients started on ramipril between 1 September 2002 to 31 August 2005 and were discontinued and switch to another antihypertensive within the switching period (<60 days for 1-month discontinuers and <120 days for 2- to 6-month discontinuers).

Switch to centrally acting antihypertensives1 (0.2)1 (0.5)0 (0.00)2 (0.3)
Switch to α-adrenoceptor blocking drugs3 (0.7)9 (4.3)2 (1.1)14 (1.7)
Switch to other angiotensin-converting enzyme inhibitors20 (4.8)18 (8.6)6 (3.4)44 (5.5)
Switch to calcium channel blockers26 (6.2)19 (9.1)10 (5.6)55 (6.8)
Switch to diuretics7 (1.7)14 (6.7)2 (1.1)36 (4.5)
Switch to β-adrenoceptor blocking drugs32 (7.6)16 (7.7)5 (2.8)53 (6.6)
Switch to angiotensin-II receptor antagonists60 (14.3)50 (23.9)36 (20.2)146 (18.1)
Total140 (33.3)a 98 (46.9)b 56 (31.5)c 294 (36.4)

Simvastatin

Approximately 10% (843) of patients discontinued simvastatin within 1 month, 3.4% (303) between 2 and 3 months and another 3.2% (283) between 4 and 6 months from the index prescription throughout the study period (Table 2). 28.8% (411) of patients who discontinued simvastatin within 6 months were switched to another lipid-regulating medicine. Only 8.1% (116) of patients were already being prescribed with another lipid-regulating therapy prior to simvastatin discontinuation. A high number of patients [63.1% (902)] had discontinued simvastatin without being prescribed any lipid-regulating medicine which translates to 10.2% (902 of 8849 patients) of all patients started on simvastatin therapy (Table 4).

Table 2. Simvastatin drug-switching information according to lipid-regulating drug class
Days on simvastatin31 days or less (843 patients) (%)32–95 days (303 patients) (%)96–185 days (283 patients) (%)185 days or less (1429 patients) (%)
  1. a

    2 were patients switched to 2 lipid-regulating drug at the same time.

  2. b

    3 were patients switched to 2 lipid-regulating drug at the same time.

  3. c

    4 were patients switched to 2 lipid-regulating drug at the same time.

  4. Patients started on simvastatin between 1 September 2002 to 31 August 2005 and discontinued and switched to another lipid-regulating drug within the switching period (<60 days for 1-month discontinuers and <120 days for 2- to 6-month discontinuers).

Switch to ezetimibe or ezetrol17 (2.0)8 (2.6)10 (3.5)35 (2.5)
Switch to fibrates4 (0.5)3 (1.0)3 (1.1)10 (0.7)
Switch to other statins165 (19.6)112 (37.0)98 (34.6)375 (26.2)
Total184 (21.8)a 120 (39.6)b 107 (37.8)c 411 (28.8)

Angiotensin-II receptor blockers (ARBs)

Angiotensin-II receptor blockers had the lowest discontinuation rate with 14% (455) of patients discontinued ARB within 6 months of the index prescription. Approximately 9% (281) of patients discontinued ARB within 1 month, 3% (93) between 2 and 3 months and another 2.5% (81) between four and 6 months throughout the study period (Table 3). 30.6% (139) of patients switched to other antihypertensive therapy after being discontinued from ARBs within 6 months of the index prescription. 39.3% (179) of patients were already being prescribed with other antihypertensives prior to ARBs discontinuation, whereas another 30.1% (137) of patients without any continuation of antihypertensive treatment (Table 4).

Table 3. Angiotensin receptor blocker (ARB) drug-switching information according to antihypertensive drug class
Days on ARBs31 days or less (281 patients) (%)32–95 days (93 patients) (%)96–185 days (81 patients) (%)185 days or less (455 patients) (%)
  1. a

    9 patients were switched to 2 antihypertensives at the same time.

  2. b

    9 patients were switched to 2 antihypertensives at the same time.

  3. c

    6 patients were switched to 2 antihypertensives at the same time.

  4. Patients started on an ARB between 1 September 2002 to 31 August 2005 and were discontinued, and switch to another antihypertensive drug within the switching period (<60 days for 1-month discontinuers and <120 days for 2- to 6-month discontinuers).

Switch to centrally acting antihypertensives1 (0.4)1 (1.1)0 (0.00)2 (0.4)
Switch to α-adrenoceptor blocking drugs11 (3.9)7 (7.5)4 (4.9)22 (4.8)
Switch to angiotensin-converting enzyme inhibitors13 (4.6)7 (7.5)9 (11.1)29 (6.4)
Switch to calcium channel blockers15 (5.3)14 (15.1)15 (18.5)44 (9.7)
Switch to diuretics25 (8.9)9 (9.7)2 (2.5)36 (7.9)
Switch to β-adrenoceptor blocking drugs13 (4.6)11 (11.8)5 (6.2)29 (6.4)
Total69 (24.6)a 40 (43.0)b 29 (35.8)c 139 (30.6%)
Table 4. Patients discontinued a drug of interest and without continuation of therapy
Patients prescribed drug of interest31 days or less32–95 days96–185 days185 days or less
  1. Determining patients without continuation of therapy.

Number of patients discontinued ramipril421 patients209 patients178 patients808 patients
Patients discontinued ramipril and were switched to another antihypertensive140 (33.3%)98 (46.9%)56 (31.5%)294 (36.4%)
Patients were already being prescribed with another antihypertensive prior to ramipril discontinuation and not switched124 (29.5%)55 (26.3%)80 (44.9%)259 (32.1%)
Patients without continuation of therapy after being discontinued from ramipril157 (37.3%)56 (26.8%)42 (23.6%)255 (31.6%)
Number of patients discontinued simvastatin843 patients303 patients283 patients1429 patients
Patients discontinued simvastatin and were switched to another lipid-regulating drug184 (21.8%)120 (39.6%)107 (37.8%)411 (28.8%)
Patients were already being prescribed with another lipid-regulating drug prior to simvastatin discontinuation and not switched70 (8.3%)21 (6.9%)25 (8.8%)116 (8.1%)
Patients without continuation of therapy after being discontinued from simvastatin589 (69.9%)162 (53.5%)151 (53.4%)902 (63.1%)
Number of patients discontinued ARB281 patients93 patients81 patients455 patients
Patients discontinued ARB and were switched to another antihypertensive69 (24.6%)40 (43.0%)29 (35.8%)139 (30.6%)
Patients were already being prescribed with another antihypertensive prior to ARB discontinuation and not switched101 (35.9%)40 (43.0%)38 (46.9%)179 (39.3%)
Patients without continuation of therapy after being discontinued from ARB111 (39.5%)13 (14.0%)14 (17.3%)137 (30.1%)

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References

Drug switching is a common medical practice. Continuation of care and the potential of using prescription data to determine drug switching were assessed in this paper. It is important for patients suffering from chronic conditions, such as hypertension and hyperlipidaemia, to receive life-long continuation of care and medication. Patients included in this study were started on therapy after being recorded as suffering from hypertension or hyperlipidaemia and therefore would require life-long medication. Although drug discontinuation leading to drug switching is likely to occur, drug discontinuation without a drug switch to an alternative therapy may indicate a complete stoppage of therapy. This complete discontinuation may represent a negative event for the treatment of chronic disease, especially for patients diagnosed with hypertension, as there is currently no cure for hypertension and the lack of therapy to control the disease will lead to an increased chance of morbidities such as stroke, ischaemic heart disease and death [15].

Identifying a complete stoppage of therapy requires details of co-prescriptions. If patients on multiple antihypertensive medications discontinue ramipril and are not switched to another antihypertensive, they will still gain protection from their other antihypertensive medication.

Ramipril

The main adverse effect reported for ramipril is cough [16]. Although ARBs have a similar mode of action, unlike ACE-Is, ARBs do not affect the metabolism of bradykinin and therefore lack the adverse effect of cough associated with ACE-Is [16]. Consequently, ARBs would be the logical ‘switch drug’ when a patient has an ADR to an ACE-I such as ramipril. The results from this study confirmed that this is indeed the case, with almost a fifth of patients who discontinued ramipril were switched to an ARB, and approximately 50% of the total patients who experienced drug switch, switched to an ARB. Approximately one-third (31.6%) of patients had a total stoppage of antihypertensive therapy after ramipril was discontinued.

Simvastatin

In terms of efficacy and safety profile, statins are the best lipid-regulating drug group [16-22]. Therefore, it is not surprising that after discontinuation of simvastatin, other statins are prescribed. After discontinuing simvastatin, 91% of the patients who experienced drug switch were switched to a different statin. Simvastatin is also the most frequently prescribed first-line agent for lipid-regulating therapy, resulting in the co-prescription of simvastatin for a low number of patients who are prescribed with other lipid-regulating medicine. However, further analysis revealed almost two-thirds (63.1%) of patients did not have any lipid-regulating therapy after simvastatin was discontinued. This figure translates to 10.2% of all patients started on simvastatin therapy who will not receive any lipid-regulating therapy after 6 months of a simvastatin index prescription. Considering the large number of patients started on simvastatin in the UK each year, this may be a cause for concern.

Angiotensin-II receptor blockers

Angiotensin-II receptor blockers are generally viewed as a group of drugs with a good safety profile, reportedly similar to placebos [23-33]. Unlike ramipril and simvastatin, which have a narrow range of switch drugs, discontinuation of an ARB was associated with a switch to a wide range of alternative antihypertensive agents. The choice of the switch drug is influenced by a number of factors, including the prescriber's experience with antihypertensive therapy, patient convenience and efficacy [1-8]. Of all the drugs of interest analysed, ARBs were associated with the smallest proportion of patients (30.1%) who completely discontinued therapy.

This study demonstrates the potential of using routinely acquired prescription data to assess drug switching, and most importantly, to assess the continuation of care for chronic medication therapy. The results obtained were consistent with expectations: patients discontinuing ramipril were switched to ARBs, patients discontinuing simvastatin were switched to other statins, and there were no specific preference for patients discontinuing ARBs. For all three study drugs, approximately 30% of patients who discontinued were switched to another agent within the fixed switching periods. However, for this study, the vital information would be the presence of the continuation of medication therapy. Approximately one-third of patients who discontinued ramipril and an ARB, and two-thirds of patients who discontinued simvastatin were without continuation of therapy. Results indicate that 4.2% of patients started on ARB, 6% of patients started on ramipril and a high of 10.2% of patients started on simvastatin will be without continuation of therapy for their chronic diseases after 6 months. This is definitely a cause for concern, especially for simvastatin, owing to the high number of patients started on simvastatin therapy each year.

Drug-switching data, especially continuation of care, are an important indicator for the standard of care. Further expansion of this system, especially if associated with specific demographics (such as age, gender, deprivation score, education profile and other variables), will enable GPs to utilize such data to decide on the best treatment option that may result in the highest compliance and thus continuation of care for their patients. This will lead to a better outcome and standard of care. Drug-switching data may also be used to identify the current preference of GPs towards a particular treatment option and switch drug. Further development to identify the reasons for a drug switch will be useful for trend analysis or prescriber behaviour analysis and cost analysis by the drug industry and health authorities. Drug-switching data may even serve as an important pharmacovigilance signal if sudden spike of drug-switching occurs.

There are a few limitations with this study. The main limitation of this study is the inability to determine the exact cause for patients being switched or not switched, thereby limiting the interpretation of results obtained and the necessary solution that can be suggested. Further development in identifying the exact reason is needed to harvest the full benefits of this methodology. It must also be recognized that all prescribing databases have inherent limitations such as incompleteness of data, data extraction errors, cross-referencing errors and absence of patient compliance data. However, these issues can be minimized with internal validation and proper scrutiny of the database as well as validation of the extracted data. Another limitation is the probability that patients might be started on an alternative therapy after the drug switch period, in which case, this information would not be included in the analysis. However, this still indicates that these patients were without any therapy for hypertension or hyperlipidaemia for a significant period of time and would most likely be considered as a negative event. Patients that obtained their antihypertensive or lipid-regulating drugs from non-PTI-registered practices while obtaining their other medications with a PTI-registered practice may cause their antihypertensive or lipid-modifying drug prescription not being detected. However, the National Health Service discourages patients registering with more than one primary practice. Owing to this and the fact that only active patients were included for analysis, the number of patients falling in this category would be minimal or negligible.

Conclusion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References

Our aim was to develop a methodology to determine and assess drug switching and continuation of care and therapy using a prescription database. Drug switching is an important indicator for the standard of care and can be an indicator for the continuation of care and therapy, as detailed. With further development of identifying the reasons for a drug switch, numerous potential uses for drug-switching information such as trend analysis by the drug industry and cost analysis by health authorities can be established. Continuous monitoring of drug switching may reveal unanticipated spike of drug switches from one drug to another drug, which may warrant further investigation by health authorities.

Hence, drug-switching data can be obtained from a prescription database to determine the continuation of care and therapy, with various other potential uses in the field of pharmacoepidemiology and drug safety.

Conflict of Interest

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References

All authors have no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Conflict of Interest
  8. References