Abstract
Aims
This research examines current hospital drug formularies (HDFs) of all Dutch general hospitals. It assesses the extent to which they recommend the same drugs, the breadth of their coverage in terms of therapeutic areas, drug groups incorporated and individuals drugs included, and their extent of conservatism by considering the year of introduction of the drugs included within groups. Furthermore, it considers the extent to which their recommendations concur and comply with those of national pharmacotherapeutic guidelines and the WHO Essential Drugs List (EDL).
Methods
Seventy-eight (81%) out of all 96 current Dutch HDFs were received of which 62 were suitable for study. Differences between HDFs and eventual associations with hospital characteristics were researched by statistical testing and case-control studies. To evaluate HDFs' concurrence with national guidelines and compliance with the WHO EDL, nine drug groups were studied in detail: benzodiazepines, calcium channel blockers, β-adrenoceptor blocking agents, ACE-inhibitors, angiotensin-II inhibitors, NSAIDs, H2-receptor antagonists, 5HT3-antagonists, and H+-pump inhibitors. Concurrence and compliance with national guidelines and the WHO EDL was defined as inclusion of recommended drugs. Non-concurrence was defined as inclusion of nonrecommended drugs.
Results
The total number of indications addressed and drug groups incorporated within HDFs varied from 28 to 72 (median 56) and from 30 to 123 (median 97), respectively. The total number of individual drug entities (pharmacological substances) included ranged from 239 to 658 (median 430) and the total number of drug products, including all different dosage forms, from 412 to 1121 (median 655). Within drug groups, drug entities first marketed were most frequently included. Teaching hospitals were most likely to include recently marketed drugs. Depending on the drug group, HDFs' concurrence and compliance with national guidelines and the WHO EDL ranged from 35% to 100%.
Conclusions
Findings indicate that Dutch HDFs are rather uniform in the indications addressed and the drug groups incorporated. However, the number of individual drug entities and drug products included within groups varies considerably. Furthermore, Dutch HDFs are considered rather conservative, as older drugs are favoured over more recent drugs. Generally, with some drug exceptions, Dutch HDFs concur and comply with recommendations in national pharmacotherapeutic guidelines and with the WHO EDL over 90%.
Keywords: 5HT3-antagonists, ACE inhibitors, angiotensin II-inhibitors, ATC code, β–adrenoceptor blocking agents, benzodiazepines, calcium channel blockers, cross – sectoral pharmacotherapy, drug expenditures, essential drugs list, H+– pump inhibitors, H2–receptor antagonists, hospital drug formularies (HDFs), ICPC – Ph subcode, NSAIDs, pharmacotherapeutic guidelines
Introduction
The population is growing and ageing. People expect equal access to healthcare, and have high expectations of (drug) therapy. Patients are using more drugs for longer periods and multiple drug use complicates clinical outcomes. As a result of new insights in pathophysiology, new and high-tech expensive products are overwhelming the pharmaceutical market. Consequently, variation in drug access by differences in drug selection and the increasing drug expenditures are cause of international concern [1–3].
Hospitals in particular have to deal with extra pressures on drug expenditures with many patients needing acute, specialized, and intensive pharmacotherapeutic treatment. Moreover, hospitals' service areas may be large and consequently include patients with a wide variety of preadmission medications. Thus, hospital pharmacy departments often stock wide ranges of (interchangeable) products. This may result in logistic problems with clinicians not thoroughly knowing all aspects of each product resulting in possible administration errors, intoxications, under-treatment, adverse effects, or drug-drug and drug–disease interactions [2, 4]. In hospitals, drug and therapeutics (D & T) committees have been traditionally responsible for rational pharmacotherapy. Rational has been defined as evidence based effective, safe, convenient, and economic [3, 5, 6]. The use of hospital drug formularies (HDFs) is considered to combine education, management, and regulation [1, 7]. At least five major types of HDFs have been described: open (comprehensive), half-open (limited), closed (restrictive), incentive-based and patient-driven [8–11]. Whereas comprehensive HDFs generally intend to include (information on) all drugs to be used, limited and restricted HDFs specifically intend to recommend and restrain drugs within drug groups or therapeutic areas, respectively. Incentive-based and patient-driven HDFs are issued by insurers and individually apply to prescribers and patients, respectively. The selection of drugs is complex but can be made more transparent with use of decision supportive matrices, such as MAUT, FDSS, Informatrix, SOJA, or SELMED [12–17]. Within D & T committees, pharmacists are key figures involved in HDF implementation, initiating and monitoring efficient pharmacotherapy within hospitals, and who subsequently have great impact on the final contents of HDFs [5, 18, 19].
Previous research assessed information on the difficulties facing Dutch hospital D & T committees, in particular with respect to the process of drug selection, and made an inventory of their formulary management techniques [19, 20]. In view of Dutch HDFs' quality assurance, further research was performed to assess detailed information on technical aspects and the organizational information included [21]. The purpose of this research was to compare and evaluate pharmacotherapeutically the contents of current limited and restrictive HDFs of Dutch general hospitals. This research assessed the extent to which HDFs recommend the same drugs, the breadth of their coverage in terms of therapeutic areas, drug groups incorporated and individuals drugs included, and their extent of conservatism by considering the year of introduction of the drugs included within groups. Furthermore, it considered the extent to which their recommendations concur and comply with those of national pharmacotherapeutic guidelines and the WHO Essential Drugs List (EDL). In view of upcoming international developments to achieve greater coherence of pharmacotherapy between general practice, hospitals, and nursing homes (primary, secondary, and tertiary sector of healthcare, respectively)[39], selected sets of indications and drug groups related to cross-sectoral pharmacotherapy (1) were studied in particular.
Methods
As part of a nation-wide survey about rational pharmacotherapy, hospital pharmacists and clinical pharmacologists of all 121 Dutch general hospitals [20] were requested to send a copy of the current HDF for research purposes, if available. In the case of nonresponse 8 weeks after the request, a written reminder was sent. All additional HDFs that were received up to 4 weeks after the reminder were included in the research.
One hospital did not provide information on the availability of an HDF. Eight hospitals did not have an HDF and 10 HDFs were used in more than one hospital. Six of the remaining 102 hospitals did not have a printed copy of the HDF. As a consequence, a total of 96 different current Dutch HDFs were eligible for research. Seventy-eight (81%) of these were received within the time limit. Response was uniform across hospital characteristics such as type (academic, nonacademic, and teaching, nonteaching), geographical region (north, south, west, east, central), type of concentration of healthcare services (rural, urban) and size (small <300 beds, medium 300–600 beds, large >600 beds). Eight HDFs were incomplete (parts missing), three were outdated (> 5-year-old or drafted before the publication date of the national pharmacotherapeutic guidelines under study), and five were invalid (unclear written corrections) and were excluded. All but nine HDFs mentioned guidance on prescribing as their main aim. The remaining nine were considered comprehensive HDFs rather than limited and restrictive HDFs and served as stock-lists or primarily focused on control of drug expenditures. To make sure that the research concentrated on HDFs seeking the same goal instead of comparing HDFs with fundamentally different goals these nine were also excluded. Consequently, the final percentage of HDFs included for research was 65 (62/96).
Criteria derived from the international literature [4, 22–24] and occupational groups were used to compare and evaluate the HDFs on the following 7 aspects: (1) the total number of indications addressed, and more specifically, (2) whether a selected set of 31 indications related to cross-sectoral pharmacotherapy was addressed (Table 1: ICPC-Ph codes 3 characters). If indications were not explicitly mentioned, these were identified based on ‘marker’ drugs, e.g. the indication angina pectoris was identified by the incorporation of nitrates, (3) the total number of drug groups, individual drug entities, and drug products, including all dosage forms, incorporated, and more specifically, (4) whether a selected set of 33 drug groups related to the selected set of indications was incorporated (Table 1: ATC code 3–5 characters), (5) the range of individual drug entities included within this selected set of drug groups; furthermore (6) identification of included individual drug entities for a selected set of nine drug groups strongly related to cross-sectoral pharmacotherapy [19]: NSAIDs, benzodiazepines, ACE inhibitors, angiotensin-II inhibitors, calcium channel blockers, β-adreno-ceptor blockers, 5HT3-antagonists, H+-pump inhibitors, and H2-receptor antagonists; (7) whether the individual drug entities included in these latter drug groups concurred with recommendations of four Dutch national pharmacotherapeutic guidelines [35–38] and whether these drug entities complied with the 10th edition of the WHO EDL. Antibiotics were excluded, as within the Netherlands these are separately incorporated in antibiotic drug formularies.
Table 1.
| Anatomical tract | HDFs | HDFs | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Disorder or disease | ICPC-Ph | n | (%) | Drug group | ATC code | n | (%) | Range*** | Median |
| General | |||||||||
| Pain | A01 | 62 | (100) | Opioids | N02A | 62 | (100) | 2–15 | 9 |
| Blood | |||||||||
| Anaemia | B80-B81 | 60 | (97) | Iron preparations | B03A | 60 | (97) | 1–6 | 3 |
| Coagulation disorders | B83 | 60 | (97) | Vitamin K antagonists | B01AA | 60 | (97) | 1–3 | 2 |
| Digestive tract | |||||||||
| Ulceric disorders | D03, D84-D87 | 62 | (100) | Antacids | A02A | 62 | (100) | 1–6 | 2 |
| H2-antagonists | A02BA | 62 | (100) | 1–3 | 1 | ||||
| H+-pump inhibitors | A02 bc | 62 | (100) | 1–2 | 1 | ||||
| Nausea | D09 | 55 | (89) | H1-antagonists | N07C | 29 | (47) | 1–3 | 1 |
| 5HT3-antagonists | A04AA | 54 | (87) | 1–3 | 1 | ||||
| Diarrhoea | D11 | 62 | (100) | Intestinal adsorbents | A07B | 62 | (100) | 1–5 | 1 |
| Constipation | D12 | 62 | (100) | Laxatives | A06A | 62 | (100) | 6–16 | 12 |
| Irritable bowel syndrome | D93 | 61 | (98) | ||||||
| Chronic enteritic disorders | D94 | 61 | (98) | Anti-inflammatory agents | A07E | 61 | (98) | 1–5 | 3 |
| Eyes and ears | |||||||||
| Glaucoma | F93 | 61 | (98) | ||||||
| Vertiginous syndromes | H82 | 56 | (90) | ||||||
| Cardiovascular tract | |||||||||
| Angina pectoris, ischaemic diseases | K74-K76 | 60 | (97) | Nitrates | C01DA | 60 | (97) | 2–3 | 3 |
| Heart failure | K77 | 61 | (98) | Cardiac glycosides | C01A | 61 | (98) | 1–2 | 1 |
| Dysrhythmia | K78-K80 | 62 | (100) | ||||||
| Hypertension | K85-K86 | 62 | (100) | Diuretics | C03 | 62 | (100) | 4–15 | 9 |
| β-blockers | C07 | 62 | (100) | 2–9 | 6 | ||||
| Calcium channel blockers | C08 | 62 | (100) | 1–7 | 5 | ||||
| ACE inhibitors | C09A | 62 | (100) | 1–5 | 3 | ||||
| Angiotensin-II inhibitors | C09C | 14 | (23) | 1 | 1 | ||||
| Haemorrhoids | K96 | 54 | (87) | ||||||
| Musculoskeletal system | |||||||||
| Rheumatic pain, Arthritic disorders | L88 | 62 | (100) | NSAIDs | M01A | 62 | (100) | 2–11 | 4 |
| Central nervous system – neurological | |||||||||
| Parkinsonism | N87 | 59 | (95) | Anti-Parkinson agents | N04 | 59 | (95) | 2–12 | 10 |
| Epileptic disorders | N88 | 62 | (100) | Antiepileptic agents | N03 | 62 | (100) | 1–13 | 9 |
| Migraine | N89 | 57 | (92) | Antimigraine agents | N02C | 57 | (92) | 1–9 | 4 |
| Central nervous system – psychiatric | |||||||||
| Anxiety, insomnia | P01-P02, P06 | 61 | (98) | Benzodiazepines Barbiturates | N05BA, N05 CD N05CA | 61 10 | (98) (16) | 1–3 1 | 1 1 |
| Depressive disorders | P03, P76 | 58 | (94) | Antidepressants | N06A | 58 | (94) | 1–14 | 7 |
| Psychotic disorders | P71-P74 | 60 | (97) | Antipsychotic agents | N05A**** | 60 | (97) | 3–25 | 16 |
| Lithium preparations | N05AN | 57 | (92) | 1–2 | 1 | ||||
| Respiratory tract | |||||||||
| Emphysema, COPD, asthma | R95-R96 | 62 | (100) | β-agonists (inhalants, systemic use) Inhalation corticosteroids Inhalation parasympatholytics Cromoglycates (inhalants) | R03A, R03C | 62 | (100) | 1–7 | 3 |
| R03BA | 62 | (100) | 1–6 | 2 | |||||
| R03BB | 61 | (98) | 1–3 | 1 | |||||
| R03 bc | 57 | (92) | 1–2 | 1 | |||||
| Endocrine system and metabolism | |||||||||
| Thyroid disorders | T85-T86 | 59 | (95) | ||||||
| Diabetes mellitus | T90 | 61 | (98) | Oral hypoglycaemics | A10B | 60 | (97) | 2–6 | 4 |
| Lipid metabolism disorders | T93 | 59 | (95) | Lipid lowering agents | C10 | 59 | (95) | 1–7 | 3 |
| Urinary tract | |||||||||
| Micturition disorders | U04 | 51 | (82) | ||||||
| Reproductive system | |||||||||
| Family planning | W11-W12 | 10 | (16) | ||||||
Selected set of 31 indications related to cross-sectoral pharmacotherapy
Selected set of 33 drug groups related to cross-sectoralpharmacotherapy
Range of individual drug entities included within the drug group
ATC code ≠ N05AN.
The four Dutch national guidelines under study have been drafted by independent, scientific authorities based on formal evidence review and evidence based medicine guideline development (2). The guidelines intend to cover the treatment of the majority of patients and are supposed to be accepted in their entirety by local physicians. Two of the guidelines are ‘disease orientated’ whereas the other two are ‘drug orientated’. With respect to the drug groups under study, the guidelines overlap and are identical. HDFs' concurrence with national guidelines was defined as inclusion of the drugs of preferred choice and exclusion of the drugs disrecommended as suggested in the national pharmacotherapeutic guidelines. HDFs' compliance with the WHO EDL was defined as inclusion of the recommended drugs included in the WHO EDL.
The software used for data input and analysis were MS Access 7.0 and SPSS 9.0. The Wilcoxon-Mann–Whitney rank test was used for comparison of medians. Pearson's correlation coefficients (r2) with P values were calculated to estimate the correlation between the number of indications addressed, the number of drug groups incorporated and the number of drug entities and drug products included on the one hand and hospital size on the other hand. Furthermore, P values were calculated by χ2 testing to estimate whether there was a statistically significant relation between year of introduction of a drug to the market and inclusion of the drug in HDFs. Pearson's correlation coefficients (r2) with P values were calculated to estimate the correlation between the year of introduction of a drug to the market and the number of HDFs that included the drug. Furthermore, case-control studies were performed to calculate odds ratios (OR) with 95% confidence intervals (95%CI) with P values (χ2 testing) considering hospitals characteristics as explanatory variables and HDF conservatism and concurrence or compliance as outcomes. ‘Conservatism’ was defined as the inclusion of ‘older’ drugs. ‘Older’ was defined as marketed >15 years ago. P values <0.05 were considered statistically significant.
All indications were classified by the International Classification of Diseases in Primary Care subcode for pharmacists (ICPC-Ph) [25]. Indications were clustered and subcodes reduced to three characters, e.g. N88.2 and N88.3 were reduced to N88 (epileptic disorders). All drugs were referred to by generic name and classified by the Anatomical Therapeutic Chemical (ATC) code reduced to 3–5 characters. The year of introduction of a drug to the pharmaceutical market was obtained from the Informatorium Medicamentorum [26]. Prior to publication all data were made anonymous.
Results
The total number of indications addressed within an HDF varied from 28 to 72 (median 56) (ICPC code 3 characters). Table 1 displays the number of HDFs addressing the selected 31 indications specifically related to cross-sectoral pharmacotherapy. The total number of drug groups incorporated within an HDF varied from 30 to 123 (median 97) (ATC code 3–5 characters). Table 1 displays the number of HDFs incorporating the selected 33 drug groups specifically related to cross-sectoral pharmacotherapy and the range and median value of individual drug entities included within these drug groups. The total number of individual drug entities included within an HDF ranged from 239 to 658 (median 430) and the total number of drug products from 412 to 1121 (median 655). The median numbers were statistically significantly higher for academic and teaching hospitals as compared with nonacademic and nonteaching hospitals, respectively (Wilcoxon-Mann–Whitney test: P = 0.05). There was no statistically significant correlation between the number of indications addressed, the number of drug groups incorporated, and the number of drug entities and products included within groups on the one hand and the hospital size on the other hand.
Identification of individual drug entities included within the 9 drug groups that were specifically studied is presented in Tables 2, 3 and 4. The drug entities most frequently included are marked in bold. Within all the groups drug entities first marketed (older drugs) were statistically significantly more frequently included than drug entities recently marketed (recent drugs) (χ2 tests; P < 0.001). For NSAIDs and β-adrenoceptor blockers, there was a statistically significant correlation between the year of introduction to the market and the number of HDFs that included the drug (r2 = −0.71, P < 0.005 and r2 = −0.62, P < 0.01, respectively). Teaching hospitals and large hospitals (> 600 beds) were more likely to include recent drugs than nonteaching and medium-sized and small hospitals (OR: 1.98; 95%CI: 1.88–2.06, P < 0.01 and OR: 1.66; 95%CI: 1.43–1.79, P < 0.01, respectively).
Table 2.
Identification of individual drug entities included within drug groups in Dutch HDFs (n = 62). The cardiovascular tract.
| HDFs | HDFs | ||||||
|---|---|---|---|---|---|---|---|
| Drug group (ATC code) | Year* | n | (%) | Drug group (ATC code) | Year* | n | (%) |
| β-adrenoceptor blockers(C07) | Calcium channel blockers (C08) | ||||||
| Nonselective (C07AA) | Selective (C08C) mainly vascular effects | ||||||
| Alprenolol *** | 1967 | 2 | (3) | Amlodipine ** | 1990 | 50 | (81) |
| Bopindolol | 1988 | 0 | (0) | Felodipine | 1987 | 7 | (11) |
| Oxprenolol | 1968 | 4 | (6) | Isradipine | 1989 | 5 | (8) |
| Pindolol | 1970 | 28 | (45) | Lacidipine | 1992 | 1 | (2) |
| Propranolol+ | 1964 | 50 | (81) | Lercanidipine *** | 1997 | 0 | (0) |
| Sotalol ** | 1974 | 61 | (98) | Nicardipine | 1986 | 10 | (16) |
| Tertatolol | 1987 | 0 | (0) | Nifedipine **/+ | 1975 | 61 | (98) |
| Timolol | 1974 | 3 | (5) | Nimodipine Nisoldipine | 1985 1990 | 37 5 | (60) (8) |
| Selective (C07AB) | Nitrendipine | 1985 | 0 | (0) | |||
| Acebutolol | 1973 | 10 | (16) | ||||
| Atenolol **/+ | 1975 | 58 | (94) | Selective (C08D) direct cardiac effects | |||
| Betaxolol *** | 1983 | 0 | (0) | Diltiazem ** | 1973 | 61 | (98) |
| Bevantolol | 1988 | 0 | (0) | Gallopamil *** | 1989 | 0 | (0) |
| Bisoprolol | 1986 | 24 | (39) | Mibefradil *** | 1997 | 0 | (0) |
| Celiprolol *** | 1987 | 8 | (13) | Verapamil **/+ | 1963 | 62 | (100) |
| Esmolol | 1990 | 23 | (37) | ||||
| Metoprolol ** | 1975 | 59 | (95) | Non-selective (C08E) | |||
| Nebivolol *** | 1995 | 0 | (0) | Bepridil | 1981 | 0 | (0) |
| Also α-adrenoceptor blocker (C07AG) | |||||||
| Carvedilol *** | 1992 | 3 | (5) | ||||
| Labetalol ** | 1977 | 58 | (94) | ||||
| ACE-inhibitors(C09A) | Angiotensin-II inhibitors(C09C) None | 48 | (77) | ||||
| Benazepril | 1991 | 0 | (0) | Candesartan | 1997 | 0 | (0) |
| Captopril **/+ | 1979 | 62 | (100) | Eprosartan | 1997 | 0 | (0) |
| Cilazapril | 1990 | 0 | (0) | Irbesartan | 1997 | 0 | (0) |
| Enalapril(ate) ** | 1984 | 60 | (97) | Losartan | 1994 | 14 | (23) |
| Fosinopril *** | 1992 | 1 | (2) | Valsartan | 1996 | 0 | (0) |
| Lisinopril ** | 1988 | 33 | (53) | ||||
| Moexipril | 1996 | 0 | (0) | ||||
| Perindopril *** | 1989 | 7 | (11) | ||||
| Quinapril *** | 1989 | 10 | (16) | ||||
| Ramipril | 1989 | 1 | (2) | ||||
| Spirapril | 1993 | 0 | (0) | ||||
| Trandolapril *** | 1993 | 0 | (0) | ||||
Year of introduction to the pharmaceutical market
Recommended by national guidelines
Not preferred or insufficiently evaluated by guideline authorities; Drug entities without asterisks are not considered in national pharmacotherapeutic guidelines
Included in the 10th edition of the WHO EDL.
Table 3.
Identification of individual drug entities included within drug groups in Dutch HDFs (n = 62). The digestive tract.
| HDFs | HDFs | ||||||
|---|---|---|---|---|---|---|---|
| Drug group (ATC code) | Year* | n | (%) | Drug group (ATC code) | Year* | n | (%) |
| H2-receptor antagonists (A02BA) | 5HT3antagonists(A04AA) | ||||||
| Cimetidine+ | 1976 | 22 | (35) | None | 8 | (13) | |
| Famotidine | 1987 | 24 | (39) | Dolasetron | 1997 | 0 | (0) |
| Nizatidine | 1988 | 0 | (0) | Ondansetron | 1990 | 11 | (18) |
| Ranitidine ** | 1981 | 56 | (90) | Granisetron ** | 1991 | 38 | (61) |
| Roxatidine *** | 1986 | 0 | (0) | Tropisetron | 1992 | 11 | (18) |
| H+ pump inhibitors(A02bc) | |||||||
| Lansoprazole*** | 1992 | 1 | (2) | ||||
| Omeprazole ** | 1988 | 61 | (98) | ||||
| Pantoprazole *** | 1995 | 0 | (0) | ||||
Year of introduction to the pharmaceutical market
Recommended by national guidelines
Not preferred or insufficiently evaluatedby guideline authorities; Drug entities without asterisks are not considered in the national pharmacotherapeutic guidelines
Included in the 10th edition of the WHO EDL.
Table 4.
Identification of individual drug entities included within drug groups in Dutch HDFs (n = 62). The musculoskeletal and central nervous system.
| HDFs | HDFs | ||||||
|---|---|---|---|---|---|---|---|
| Drug group (ATC code) | Year* | n | (%) | Drug group(ATC code) | Year* | n | (%) |
| NSAIDs(M01A/N02BB/N02BG) | Benzodiazepines(N05BA/N05 CD) None | 1 | (2) | ||||
| Acetic acid derivates | Alprazolam *** | 1983 | 0 | (0) | |||
| Aceclofenac | 1997 | 0 | (0) | Bromazepam | 1974 | 0 | (0) |
| Diclofenac ** | 1974 | 60 | (97) | Brotizolam | 1982 | 2 | (3) |
| Indomethacin | 1963 | 56 | (90) | Chlorodiazepoxide *** | 1960 | 0 | (0) |
| Sulindac | 1976 | 11 | (18) | Clobazam *** | 1975 | 0 | (0) |
| Tolmetine | 1975 | 2 | (3) | Clorazepinic acid *** | 1967 | 0 | (0) |
| Diazepam **/+ | 1961 | 23 | (37) | ||||
| Oxicam derivates | Flunitrazepam *** | 1974 | 11 | (18) | |||
| Meloxicam | 1996 | 1 | (2) | Flurazepam *** | 1968 | 19 | (31) |
| Piroxicam | 1979 | 27 | (44) | Ketazolam | 1980 | 0 | (0) |
| Tenoxicam *** | 1986 | 0 | (0) | Loprazolam | 1983 | 0 | (0) |
| Loprazolam | 1971 | 17 | (28) | ||||
| Proprionic acid derivates | Lormetazepam | 1980 | 25 | (40) | |||
| Fenoprofen | 1974 | 0 | (0) | Medazepam *** | 1968 | 27 | (44) |
| Flurbiprofen | 1977 | 0 | (0) | Midazolam | 1982 | 0 | (0) |
| Ibuprofen **/+ | 1969 | 59 | (95) | Nitrazepam ** | 1965 | 47 | (76) |
| Ketoprofen | 1973 | 19 | (31) | Nordazepam *** | 1975 | 0 | (0) |
| Naproxen ** | 1973 | 57 | (92) | Oxazepam | 1964 | 19 | (31) |
| Tiaprofenic acid | 1980 | 0 | (0) | Prazepam *** | 1973 | 0 | (0) |
| Temazepam ** | 1969 | 61 | (98) | ||||
| Pyrazolinon derivates | Triazolam | 1978 | 0 | (0) | |||
| Aminophenazone | 1887 | 0 | (0) | Zolpidem | 1988 | 4 | (7) |
| Phenazone | 1884 | 0 | (0) | Zopiclon | 1985 | 12 | (19) |
| Metamizol *** | 1946 | 3 | (5) | ||||
| Propyphenazone | 1951 | 0 | (0) | ||||
| Other | |||||||
| Azapropazone | 1970 | 9 | (14) | ||||
| Floctafenine | 1976 | 0 | (0) | ||||
| Ketorolac | 1990 | 0 | (0) | ||||
| Phenylbutazone | 1949 | 13 | (21) | ||||
| Nabumeton | 1985 | 7 | (11) | ||||
| Proquazone | 1977 | 0 | (0) | ||||
| Tolphenaminic acid | 1975 | 0 | (0) | ||||
Year of introduction to the pharmaceutical market
Recommended by national guidelines
Not preferred or insufficiently evaluated by guideline authorities; Drug entities without asterisks are not considered in the national pharmacotherapeutic guidelines
Included in the 10th edition of the WHO EDL.
Depending on the drug group, HDFs' concurrence with national pharmacotherapeutic treatment guidelines varied (Table 2,3,4). With respect to β-adrenoceptor blockers, most HDFs concurred with national guidelines preferring inclusion of at least one of each type (nonselective, selective, and α-and β-adrenoceptor blocking properties). More specifically, metoprolol, sotalol, atenolol, and labetolol are preferred, as these four cover all cardiovascular indications. Metoprolol is also effective in the treatment of migraine and labetolol can be used during pregnancy. Nebivolol and betaxolol are considered to be expensive and as yet insufficiently monitored by guideline authorities. As a consequence, these were not preferred and accordingly not included in any HDF. However, celiprolol, carvedilol, and alprenolol are considered to be of little added value and are therefore not preferred. Nevertheless, some HDFs included these agents. With respect to calcium channel blockers, HDFs also concurred with national guidelines preferring nifedipine, dilitiazem, amlodipine, and verapamil. Accordingly, most HDFs included these. Moreover, lercanidipine, gallopamil, and mibefranil are not preferred or as yet insufficiently monitored and accordingly were not included in any HDF. Nonetheless, a majority of HDFs included short-acting nifedipine capsules that are considered clinically ineffective. With respect to ACE-inhibitors, HDFs' concurrence with national guidelines differed. Guidelines prefer enalapril(ate), captopril, and lisinopril. Accordingly, the former two were most frequently included. However, lisinopril was included in 35% of all HDFs. Furthermore, inclusion of fosinopril, perindopril, quinapril and trandolapril is discouraged as these are considered to be of little added value and also very expensive. However, some HDFs still included these drugs. With respect to angiotensin-II-inhibitors, evaluation of HDFs' concurrence was not possible. Most angiotensin-II-inhibitors have been introduced very recently. Consequently, guideline authorities have not as yet fully evaluated these agents.
With respect to H2-receptor antagonists, HDFs concurred with the national guidelines. Although some guidelines tend to advise ranitidine, in general, all drugs within this group are considered to be equally preferred. Ranitidine was included most frequently because it is perceived to have less drug–drug interactions but many HDFs included cimetidine and famotidine as well. Guidelines suggest this selection should be based on costs. Consequently, roxatidine is not recommended. With respect to H+-pump inhibitors, national guidelines favour omeprazole over lansoprazole and pantoprazole. The latter two are considered to be of little added value. Accordingly, most HDFs only included omeprazole. With respect to 5HT3-receptor antagonists, most HDFs concurred with national guidelines preferring ondansetron and granisetron. Some HDFs have not incorporated 5HT3-receptor antagonists at all as consensus has not as yet fully been achieved within the hospital D & T committees. Ondansetron may be favoured theoretically over granisetron because the latter is administered parenterally and can not be used at home. However, granisetron is still included more frequently.
HDF compliance with national guidelines is very strict with respect to NSAIDs. Ibuprofen, naproxen, and diclofenac are those first marketed and recommended by the guidelines. These drugs can be used by children, elderly and pregnant women. Accordingly, these are included in most HDFs. Some additionally include indomethacin, which may have some additional value in severe rheumatoid arthritis. Guidelines strongly discourage the inclusion of tenoxicam and metamizol. Nevertheless, the latter is included in some HDFs.
With respect to benzodiazepines, HDFs' concurrence was high. Flunitrazepam, alprazolam, chlorodiazepoxide, clobazam, clorazepinic acid, nordazepam and prazepam are not recommended due to their adverse effects and as a result, these were not included in any HDF. However, flurazepam and medazepam are also nonrecommended but were still included in many HDFs as some patients use this drug as part of their preadmission medication. Guidelines recommend the short acting temazepam and the long-acting diazepam and nitrazepam, which, accordingly, most HDFs included.
In summary, concurrence with guidelines ranged from 53% (inclusion of lisinopril) to 100% (inclusion of verapamil) and nonconcurrence rose up to 44% (inclusion of medazepam). A similar range was observed for compliance with the 10th edition of the WHO EDL. Compliance ranged from 35% (inclusion of cimetidine) to 100% (inclusion of captopril). No statistically significant association was found between hospital characteristics and (non) concurrence or compliance.
Discussion
These findings indicate that the pharmacotherapeutic quality of Dutch HDFs is rather uniform in the indications addressed and the drug groups incorporated related to cross-sectoral pharmacotherapy. Furthermore, depending on the drug group, the individual drug entities included concur and comply differently with Dutch national pharmacotherapeutic guidelines and the WHO EDL [35–38]. Nevertheless, in general, most HDFs include over 90% of the recommend drugs. Also, Dutch HDFs are considered rather conservative, as within drug groups, drug entities that were marketed first are favoured over recently marketed drugs. The total number of drug entities and drug products included varies considerably with some HDFs including wide ranges within specific drug groups. This may, in view of HDFs' quality assurance and control of drug expenditures, be questioned, as it appears contrary to the concept of limited and restrictive HDFs.
Although findings have shown that most HDFs address and incorporate the selected sets of indications and drug groups related to cross-sectoral pharmacotherapy, none specifically clarifies why some are addressed or incorporated and others not. In general, it is suggested that not all indications need to be addressed to make HDFs efficient [27]. Particularly those indications that implicate expensive pharmacotherapy or have a high prevalence or incidence rate, based on morbidity registration, should be addressed [23]. Unlike Rucker [4], we also believe that not all drug groups need to be incorporated. We suggest that those that account for a high proportion of drug expenditures, those related to cross-sectoral pharmacotherapy (in view of seamless care), those that require careful therapeutic drug monitoring (toxicity), and those that may cause drug–drug interactions or severe adverse effects at least need to be incorporated [19].
Furthermore, our findings on wide ranges and high median numbers correspond with previous research that identified certain drug groups as difficult to achieve consensus within hospital D & T committees [19, 22]. Within these groups a large number of drug entities is available on the pharmaceutical market, with only slight differences regarding clinical effectiveness, adverse effects, and price. Therefore, preferences for particular drug entities may strongly depend on clinicians' involvement in clinical trials and clinical experience. Marketing strategies aimed at specific clinicians (cardiologist, psychiatrist, and internists) and personal benefits are said to obstruct any objective achievement of HDF consensus. With respect to these groups, national pharmacotherapeutic guidelines fail to present clear recommendations about which drug entities are favoured over others. Moreover drug entities within these groups may be used off-label and one drug entity is frequently not available in all dosage forms. This latter implies that different drug entities are selected to include all forms. Finally, individual drug entities, in particular those that act on the central nervous system, may have slight but essential different clinical effects in different patients. As a consequence, psychiatrists are rather reluctant to make a selection. Thürmann et al. suggest that wide ranges indicate an unaccomplished rationalization process within D & T committees [22].
Generally, the number of drug entities and drug products included is comparable with HDFs in other countries [22, 24, 28]. An increase is likely since it not only depends on the rationalization in the process of drug selection but also on the growing diversity of patients' preadmission medications, the continuous introduction of new dosage forms, and the expansion of registered indications per individual drug entity. In view of pharmacotherapeutic quality assurance, it is suggested that within HDFs, a switch from one drug entity or drug product to another within a specific group, based on costs, within a period of 2 years is undesirable for it enhances confusion and may implicate medication errors [28, 29].
Within the drug groups under study, older drugs are favoured over more recently introduced drugs. This also supports previous research which indicates that within Dutch hospital D & T committees, evidence based literature and long clinical experience outweigh considerations of potential new drug entities which may have only slightly added value. In contrast, teaching and large hospitals, often involved in clinical trials, more frequently included recently marketed drug entities because of good experience in specific patients [19].
Furthermore, findings on concurrence and compliance are similar to those from other countries [22]. In general, each HDF is as a product of local opinion and consensus about pharmacotherapeutic treatments. Inclusion of drugs not preferred by guidelines or evaluated insufficiently by guideline authorities does not indicate inappropriateness. For example, inclusion of losartan may be a pragmatic response to clinicians' demands. Absence of guidelines demonstrates the limitations of development of such guidelines. Also, continuing inclusion of medazepam may be a pragmatic decision because old prescribing habits die hard. Moreover, responding to up-to-date evidence, nonconcurrent HDFs may well be ahead of national guidelines as Dutch national guidelines are often criticised for conservatism themselves.
Some may consider no need for superstructures that compare and evaluate local HDFs. However, we suggest that in view of quality assurance and equal access to healthcare, external independent referees, representing medical and pharmaceutical science, medical and pharmaceutical practice, government, and eventually healthcare insurers, should evaluate HDFs pharmaco- therapeutically. In view of healthcare economics, governments are stressing greater pharmacotherapeutic coherence between hospitals and between hospitals and other sectors of healthcare [4]. Finally, we emphasise the key role of technical aspects and organizational information included. Rational drug selection is essential but without proper presentation and supportive therapeutic guidance information to guarantee appropriate treatment, HDFs may not have any value in daily clinical practice [4, 21, 23, 30–34].
Acknowledgments
The authors gratefully acknowledge the assistance of C. J. De Blaey, PhD., MPharmSc., RPh., Scientific Institute Dutch Pharmacists, The Hague, The Netherlands, C. S. De Vries, PhD., MPharmSc., RPh., Medicines Monitoring Unit, Department of Clinical Pharmacology, University of Dundee, United Kingdom, and L.G. Pont, MPharmSc., RPh., Department of Clinical Pharmacology, University of Groningen, The Netherlands, in research design and manuscript preparation. The authors express their thanks to all pharmacists who supplied the HDFs.
Footnotes
(1) Cross-sectoral pharmacotherapy comprises drugs that are not restricted to a particular sector of Dutch healthcare but that are used by patients across the interface of sectors. Therefore, these drugs are prescribed by general practitioners, hospital clinicians, and clinicians in nursing homes. In particular, these drug were previously identified as difficult to achieve consensus about within hospital D&T committees [19]
(2) Dutch national pharmacotherapeutic guidelines: ‘NHG-standaarden’.;- drafted by the Royal Dutch Society for Advancement of Medicine [36]. ‘CBO-consensi’; drafted by the Central Guidance Authority for Intercollegiate Examination of Medical Specialists [37]. ‘Geneesmiddelenbulletin’. National Drug and Therapeutics Bulletin/Monthly Medical Letter drafted by the Drug Bulletin Institute [38]. ‘Farmacotherapeutisch Kompas’. National Formulary/Annual Drugs Index drafted by the Dutch Health Insurance Fund Council [35].
References
- 1.Hogerzeil HV. Promoting rational prescribing: an international perspective. Br J Clin Pharmacol. 1995;39:1–6. doi: 10.1111/j.1365-2125.1995.tb04402.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Martinez Bengoechea MJ, Messori A, Berto V, Becagli P, Font M, Martini N. Hospital formulary and drug selection. Eur Hosp Pharm. 1997;3:89–90. [Google Scholar]
- 3.Gilley J. Towards rational prescribing. Better prescribing takes time. Br Med J. 1997;308:731–732. [Google Scholar]
- 4.Rucker TD. Quality control of hospital formularies. Pharm World Sci. 1988;10:145–150. doi: 10.1007/BF01959422. [DOI] [PubMed] [Google Scholar]
- 5.Krämer I. Rationalization of drug therapy by the hospital pharmacist. Int J Clin Pharmacol Ther. 1995;33:473–474. [Google Scholar]
- 6.Bochner F, Martin ED, Burgess NG, Somogyi AA, Misan GMH. Controversies in treatment. How can hospitals ration drugs? Br Med J. 1994;308:901–908. doi: 10.1136/bmj.308.6933.901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Glaeske G. Quality control of drug prescriptions by positive lists – The European Formulary. Int J Clin Pharmacol Ther. 1994;32:403–408. [PubMed] [Google Scholar]
- 8.ASHP. ASHP national survey. Am J Hosp Pharm. 1993;50:1394–1395. [PubMed] [Google Scholar]
- 9.Giaquinta D. Drug formularies – good or evil? A view from a managed care provider. Cardiol. 1994;85(Suppl 1):30–35. doi: 10.1159/000176755. [DOI] [PubMed] [Google Scholar]
- 10.Hanson EC, Shepherd M. Formulary restrictiveness in health maintenance organizations. J Soc Adm Pharm. 1994;11:54–56. [Google Scholar]
- 11.Woodhouse KW. Drug formularies – good or evil? The clinical perspective. Cardiol. 1994;85(Suppl 1):36–40. doi: 10.1159/000176756. [DOI] [PubMed] [Google Scholar]
- 12.Vermeij DJB. The general practitioner's drug formulary. A model for implementation of common options. J Drug Res. 1988;13:116–119. [Google Scholar]
- 13.Janknegt R, Steenhoek A. The System of Objectified Judgement Analysis (SOJA) Drugs. 1997;53:550–562. doi: 10.2165/00003495-199753040-00002. [DOI] [PubMed] [Google Scholar]
- 14.Schumacher GE. Multiattribute evaluation in formulary decison making as applied to calcium-channel blockers. Am J Hosp Pharm. 1991;48:301–308. [PubMed] [Google Scholar]
- 15.Brenninkmeijer RF, Vermeij DJB, Hes R, Van der Kleijn E. InforMatrix. Pharmaceutisch Weekblad. 1994;129:1185–1190. [Google Scholar]
- 16.Senthilkumaran K, Shatz SM, Kalies RF. Computer-based support system for formulary decisions. Am J Health Syst Pharm. 1987;44:1362–1366. [PubMed] [Google Scholar]
- 17.Janknegt R, Van der Kuy A, Declerck G, Idzikowsky C Hypnotics. Drug selection by means of the SOJA method. Pharmacoeconomics. 1996;10:152–163. doi: 10.2165/00019053-199610020-00007. [DOI] [PubMed] [Google Scholar]
- 18.Levens-lipton H, Byrns PJ, Soumerai SB, Chrischilles EA. Pharmacists as agents of change for rational drug therapy. Int J Technol Ass Health Care. 1995;11:484–508. doi: 10.1017/s0266462300008692. [DOI] [PubMed] [Google Scholar]
- 19.Fijn R, Brouwers JRBJ, Engels SAG, Knaap RJ, De Jong-van den Berg LTW. Survey of Dutch hospital Drug and Therapeutics (D & T) committees. Br J Clin Pharmacol. 1999;48:239–246. doi: 10.1046/j.1365-2125.1999.00001.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fijn R, De Jong-van den Berg LTW, Brouwers JRBJ. Rational pharmacotherapy in The Netherlands: Formulary management in Dutch hospitals. Pharm World Sci. 1999;21:74–79. doi: 10.1023/a:1008654609916. [DOI] [PubMed] [Google Scholar]
- 21.Fijn R, De Vries CS, Engels SAG, Brouwers JRBJ, De Blaey CJ, De Jong-van den Berg LTW. The quality of Dutch hospital drug formularies. An evaluation of technical features and organisational information. Pharm World Sci. 1999;21:101–106. doi: 10.1023/a:1008680021854. [DOI] [PubMed] [Google Scholar]
- 22.Thürmann PA, Harder S, Steioff A. Structure and activities of hospital drug committees in Germany. Eur J Clin Pharmacol. 1997;52:429–435. doi: 10.1007/s002280050315. [DOI] [PubMed] [Google Scholar]
- 23.Kamps GB, Meyboom-de Jong B. Evaluation of regional formularies for general practitioners. Nederlands Tijdschrift Voor Geneeskunde. 1997;141:1002–1007. [PubMed] [Google Scholar]
- 24.Rucker TD. Superior hospital formularies: a critical analysis. Hosp Pharm. 1982;17:465–524. [PubMed] [Google Scholar]
- 25.Van Mil JWF, Brenninkmeijer R, Tromp TFJ. The ICPC coding system in pharmacy: developing a subset, ICPC-Ph. Pharm World Sci. 1998;20:38–42. doi: 10.1023/a:1008631413768. [DOI] [PubMed] [Google Scholar]
- 26.Anonymous. Informatorium Medicamentorum. Dutch National Formulary. 28. Gravenhage: Royal Dutch Society for the Advancement of Pharmacy; 1998. [Google Scholar]
- 27.Anonymous. Constructing a practice formulary: a learning experience. Drug Ther Bull. 1991;29:25–26. [PubMed] [Google Scholar]
- 28.Anonymous. Die Arzneimittelkomission im Krankenhaus – ein Beitrag zur rationalen und kostenbewussten Therapie [in German] Der Arzneimittelbrief. 1998;21:33–35. [Google Scholar]
- 29.Sprague KL. Formulary design & management: Today's weapon to shatter cost and improve healthcare quality. Hosp Form. 1993;28:429–435. [Google Scholar]
- 30.Plumridge RJ, Stoelwinder JU, Berbatis CG. Improving patient care and pharmacy management: the effects of hospital formularies. Drug Intell Clin Pharm. 1984;18:652–656. doi: 10.1177/106002808401800732. [DOI] [PubMed] [Google Scholar]
- 31.Steenhoek A, Van Soest MM. Drug formularies in hospitals [in Dutch] Pharmaceutisch Weekblad. 1998;133(26):970–973. [Google Scholar]
- 32.Sutters CA. The management of a hospital formulary. J Clin Pharm Ther. 1990;15:59–76. doi: 10.1111/j.1365-2710.1990.tb00357.x. [DOI] [PubMed] [Google Scholar]
- 33.Tugwell AC, Thurston DR, Barret CW. Design and preparation of a formulary – Guide to the prescribing of medicines. J Clin Hosp Pharm. 1984;9:311–319. doi: 10.1111/j.1365-2710.1984.tb01092.x. [DOI] [PubMed] [Google Scholar]
- 34.Branch RA, Johnston PE, Koestner JA, Bluhm R, Stratton CW, Knight JR. The formulary: an educational tool for clinical pharmacology. Clin Pharmacol Ther. 1992;51:481–487. doi: 10.1038/clpt.1992.51. [DOI] [PubMed] [Google Scholar]
- 35.Van der Kuy A, Van Luijn JCF, Van Loenhout JWA, et al. Farmacotherapeutisch Kompas 1990-98. 15. Amstelveen: Health Insurance Fund Council; 1998. [Google Scholar]
- 36.Rutten GEHM, Thomas S. Utrecht: Wetenschappelijke Uitgeverij Bunge; 1996. NHG-Standaarden 1990-96. [Google Scholar]
- 37.CBO. Utrecht: Centraal Bureau voor Intercollegiale Toetsing; 1998. CBO richtlijnen 1990-98. [Google Scholar]
- 38.Geneesmiddelenbulletin Instituut. Utrecht: Geneesmiddelenbulletin Instituut; 1998. Geneesmiddelenbulletin jaargangen 1990-98. [Google Scholar]
- 39.Fijn R, Brouwers JRBJ, De Jong-van den Berg LTW. Cross-sectoral pharmacotherapeutic coherence in the Netherlands. Int J Pharm Pract. 1999;7:159–166. [Google Scholar]