Sequential antibiotic therapy : Effective cost management and patient care

Division of Infectious Diseases, McMaster University, Hamilton, Ontario; Department of Microbiology, Université Laval and Labratoire et service d’infectiologie, Centre Hospitalier de l’Université Laval, Québec, Québec; University of British Columbia and Division of Infectious Diseases, University Hospital, Vancouver, British Columbia; Faculty of Pharmaceutical Sciences, University of British Columbia and Department of Pharmacy, Department of Medicine, Division of Infectious Diseases, Vancouver Hospital and Health Sciences Centre, Vancouver, British Columbia; Departments of Microbiology and Medicine, University of Toronto and Department of Microbiology, Mount Sinai Hospital and Princess Margaret Hospital, Toronto, Ontario; Dalhousie University and Department of Medicine, Victoria General Hospital, Halifax, Nova Scotia; Internal Medicine and Medical Microbiology, University of Manitoba and Infection Control Unit, Health Sciences Centre, Winnipeg, Manitoba Correspondence: Dr Lionel A Mandell, Head, Division of Infectious Diseases, McMaster Medical Unit, Henderson General Hospital, 711 Concession Street, Hamilton, Ontario L8V 1C3. Telephone 905-574-8520, fax 905-575-7320, e-mail lmandell@fhs.mcmaster.ca Received for publication May 17, 1995. Accepted August 11, 1995 LA MAN DELL, MG BERG ERON, MJ GRIB BLE, et al. Se quen tial an ti bi otic ther apy: Ef fec tive cost man age ment and pa tient care. Can J In fect Dis 1995;6(6):306315. The es ca lat ing costs as so ci ated with an ti mi cro bial che mo ther apy have be come of in creas ing con cern to phy si cians, phar ma cists and pa tients alike. A number of strate gies have been de vel oped to ad dress this prob lem. This ar ti cle fo cuses spe cifi cally on se quen tial an ti bi otic ther apy (SAT), which is the strat egy of con vert ing pa tients from in tra ve nous to oral medi ca tion re gard less of whether the same or a dif fer ent class of drug is used. Ad van tages of SAT in clude eco nomic bene fits, pa tient bene fits and bene fits to the health care pro vider. Po ten tial dis ad van tages are cost to the con sumer and the risk of thera peu tic fail ure. A criti cal re view of the pub lished lit era ture shows that evi dence from ran dom ized con trolled tri als sup ports the role of SAT. How ever, it is also clear that fur ther stud ies are nec es sary to de ter mine the op ti mal time for in tra ve nous to oral changeo ver and to iden tify the vari ables that may in ter fere with the use of oral drugs. Pro ce dures nec es sary for the im ple men ta tion of a SAT pro gram in the hos pi tal set ting are also dis cussed.

O VER THE PAST TWO DEC ADES, OUR THERA PEU TIC ARMAMEN - tarium has been greatly ex panded with agents that provide ex cel lent an ti mi cro bial ac tiv ity com bined with im proved phar ma coki netic fea tures and im proved ad verse ef fect profiles.At the same time, how ever, eco nomic con straints and hos pi tal bed clo sures have forced us to con sider ap proaches other than tra di tional in-hospital in tra ve nous an ti bi otic use.We now have a number of op tions that fa cili tate and sim plify pa tient care and re duce costs.These in clude home in tra venous ther apy and oral ther apy us ing agents that do not compro mise thera peu tic ef fi cacy.
This pa per re views some of the eco nomic is sues as so ciated with an ti mi cro bial ther apy and ex am ines the treat ment of in fec tions with in tra ve nous an ti mi cro bi als and the sub se quent use of oral agents.This prac tice is re ferred to as se quen tial an ti bi otic ther apy (SAT); sim ply stated, it is the prac tice of chang ing from in tra ve nous to oral dos age forms as early during a course of an ti bi otic treat ment of in fec tion as is clini cally pos si ble.

COST CON TAIN MENT IS SUES
An ti bi ot ics are among the most com monly pre scribed drugs in Can ada.Ac cord ing to the Fifth An nual Re port of the Pat ented Medi cine Prices Re view Board (1), to tal Ca na dian pat ented pre scrip tion drug reve nue was over $2 bil lion in 1992.Anti-infective agents, cost ing $354 mil lion, ac counted for the larg est number of pat ented drug prod ucts sold in Canada.The cost of in jecta ble an ti bi ot ics alone was $110 mil lion (2).In any Ca na dian hos pi tal, an ti bi ot ics of ten rep re sent the sin gle larg est com po nent of the hos pi tal phar macy budget, ac count ing for 20 to 40% of to tal drug costs.
In an at tempt to cope with in creas ing costs, sev eral strategies have been de vel oped and im ple mented and may be clas si fied as 'ed uc ative and per sua sive', 'f acil it ative' and 'restri ctive' strate gies (3).The fol low ing list in cludes most of these: • pre scriber edu ca tion • for mu lary re stric tion and re served an ti mi cro bial program • se lec tive re port ing of sus cep ti bil ity test ing • auto matic stop poli cies • thera peu tic in ter change pro grams • an ti mi cro bial or der forms • re quired con sul ta tion and phy si cian or serv ice re striction It is be yond the scope of this ar ti cle to deal with all these cost con tain ment strate gies; our pur pose is to fo cus spe cifically on SAT.

SE QUEN TIAL AN TI BI OTIC THER APY
SAT re fers to the prac tice of lim it ing the use of in tra ve nous an ti bi ot ics to the early stages of in fec tion and then con vert ing to oral agents for the du ra tion of treat ment.
SAT is not as new an ap proach to cost ef fec tive an ti bi otic pre scrib ing as one might think.Stud ies pub lished in the 1970s in volv ing chil dren with os teo mye li tis and sep tic ar thri tis dem -on strated the ef fi cacy and safety of ini tial treat ment with parenteral an ti bi ot ics fol lowed by con ver sion to oral agents as soon as the acute signs and symp toms of in fec tion were controlled (4)(5)(6)(7)(8).
In a 1991 sur vey of an ti bi otic de ci sion mak ing at a 1000bed ter ti ary care hos pi tal, Quin tili ani et al (9) re ported that anti bi otic ther apy could be 'strea mlined' in ap proxi mately 75% of pa tients in one of three ways: first, by chang ing from com bi nation ther apy to mono ther apy; sec ond, by chang ing to an other agent with a nar rower an ti mi cro bial spec trum of ac tiv ity and/or to one with pref er able phar ma coki net ics; and third, by chang ing the route of ad mini stra tion from in tra ve nous to in tramus cu lar or oral.
This con cept of stream lin ing from a more com plex, of ten more ex pen sive, regi men to a less com plex one has also been re ferred to in the lit era ture as 'ste pdown ther apy', 'switching' or 's eque ntial ther apy'.How ever, these terms have not al ways been used syn ony mously.
For ex am ple, step down ther apy and switch ing have been used to de note not only a change from in tra ve nous to the oral form of the same drug, but also a change to a dif fer ent drug once the con ver sion to oral ther apy has been made (10,11).Like wise, se quen tial ther apy has been used to de note a change from in tra ve nous to oral forms of the same drug or differ ent drugs (12,13).Stream lin ing has also been used to denote the change from com bi na tion ther apy to mono ther apy (14).It might, there fore, avoid con fu sion by es tab lish ing a com mon ter mi nol ogy for de scrib ing con ver sion from in tra venous to oral medi ca tions.Not only would this make dis cus sion of such treat ment strate gies eas ier, it would aid in cata logu ing the medi cal lit era ture deal ing with this sub ject.We pro pose that the term 's equen cing' or 's eque ntial an ti bi otic ther apy' be used to de fine the strat egy of con vert ing pa tients from in tra venous to oral medi ca tions, re gard less of whether the same or a dif fer ent class of drug is used.

AD VAN TAGES AND DIS AD VAN TAGES OF SAT
There are sound clini cal and fi nan cial rea sons to pur sue a thera peu tic strat egy that in cor po rates ap propri ate early conver sion from in tra ve nous to oral anti microbials.Some phy sicians re main re luc tant to do this, per haps be cause of a lack of knowl edge and ap pre cia tion of the ef fi cacy and ad van tages of such a strat egy, as well as the sense of se cu rity pro vided by the se rum and tis sue drug lev els ob tained us ing in tra venous drugs.This re luc tance has, un for tu nately, fos tered the un nec es sar ily pro longed use of in tra ve nous medi ca tions in the hos pi tal set ting for treat ment of in fec tions that could benefit from a short ened course of in tra ve nous treat ment fol lowed by oral ther apy.Ad van tages of SAT: Three main bene fits are as so ci ated with the use of SAT: eco nomic bene fits, pa tient bene fits and benefits to the health care pro vider.Eco nomic bene fits: With any drug, there are two costs to con sider: the more ob vi ous or ap par ent cost of the agent, referred to as the ac qui si tion cost, and the sec on dary costs related to de liv ery or ad mini stra tion of the drug.The lat ter in clude a va ri ety of fac tors, such as a phar ma cist's prepa ra tion CAN J IN FECT DIS VOL 6 NO 6 NO VEM BER/DE CEM BER 1995 and de liv ery time; a nurse's ad mini stra tion time; an cil lary supplies, such as in tra ve nous bags and tub ing; loss due to wastage; and the need for labo ra tory moni tor ing of drug se rum lev els, as is re quired with ami no gly cosides and van co my cin.All these fac tors con trib ute to the cost of drug ther apy and should be con sid ered when the cost of a par ticu lar drug is evalu ated.
Prepa ra tion and ad mini stra tion costs vary among hos pitals.Rush (15) re ported that the cost of prepa ra tion and admini stra tion added US$7.00 to the cost of each dose of in tra ve nous an ti bi otic.In Aus tra lia, such costs add be tween AUS$4.55 and $10.58 to the cost of each in tra ve nous dose (16).Oth ers have re ported that, by sim ply re plac ing one in trave nous an ti bi otic with an other that is given less fre quently, sub stan tial cost sav ings are re al ized, even when the ac qui sition cost of the re place ment drug is higher (17,18).For ex ample, at one Ameri can hos pi tal, the to tal daily ad mini stra tion cost of peni cil lin G 3x10 6 U given in tra ve nously every 4 h is US$29.26(17), yet the cost of ce fa zolin 1 g in tra ve nously every 8 h is only US$14.60.
Gen er ally, ac qui si tion costs for the in tra ve nous form of a drug are greater than those of the oral form.Some ex am ples of these com para tive costs are given in Ta ble 1.The use of oral agents ob vi ates the need for most, if not all, an cil lary costs.
At the Van cou ver Hos pi tal and Health Sci ences Cen tre, an in tra ve nous to oral an ti bi otic con ver sion pro gram has been in place since 1987.Chang ing from in tra ve nous to oral ther apy us ing drugs such as met roni da zole, clin da my cin, cipro floxacin, flu cona zole, ce fu rox ime and ce fixime has re sulted in savings of at least $30,000 yearly (10,11).Hart ford Hos pi tal in Con necti cut pro jected an an nual sav ing of US$107,637 based on an an ti bi otic stream lin ing pro gram (14).
An other ma jor fi nan cial sav ing is re al ized by the ear lier discharge of the pa tient from the hos pi tal.This elimi nates the expenses as so ci ated with hous ing a pa tient in an acute care fa cil ity solely for the pur pose of ad min is ter ing in tra ve nous anti bi ot ics (19).
The mag ni tude of cost sav ings that can ac crue by sequenc ing from in tra ve nous to oral treat ment is ap par ent from a mul ti cen tre study in the United States in volv ing 766 hos pital ized pa tients.Af ter suc cess ful con ver sion to oral ciprofloxa cin from a va ri ety of par enteral an ti mi cro bial agents, 418 pa tients were dis charged from the hos pi tal ear lier than would have oth er wise oc curred.An es ti mated to tal of 2266 hos pi tal days were saved, re sult ing in sav ings of US$793,100.Projected sav ings for to tal drug plus hos pi tali za tion costs were US$980,246 (20).Pa tient bene fits: Al though less quan ti fi able than the more tan gi ble fi nan cial gains, patient-related bene fits are nev er theless real and im por tant.In hos pi tal, the use of oral in stead of in tra ve nous drugs in creases pa tient com fort and mo bil ity, the lat ter be ing a par ticu larly im por tant con sid era tion in the elderly.By not us ing in tra ve nous lines, there is less risk of phle bitis and line-related in fec tions.The ear lier dis charge from hos pi tal also de creases the risk of de vel op ment of other noso comial in fec tions.An ear lier re turn to fam ily and, pos si bly, to work pro vides bene fits in terms of en hanced qual ity of life as well as pos si ble eco nomic bene fits.Health care pro vider bene fits: Con sid er able time is spent by phar ma cists and nurses in pre par ing and ad min is ter ing in trave nous an ti bi ot ics.Use of SAT de creases the amount of person nel time as so ci ated with drug de liv ery and, al though the ac tual bene fits of 'saved time' are dif fi cult to as sess, it may serve to free the in di vid ual for other tasks that may im prove pa tient care and en hance job sat is fac tion for the health care pro vider.

Dis ad van tages of SAT:
There are two per ceived dis ad vantages with SAT: eco nomic and risk of thera peu tic fail ure.Eco nomic dis ad van tages: While re ceiv ing medi ca tion in the hos pi tal, the pa tient is not re spon si ble for any of the drugrelated costs.How ever, once dis charged, the con sumer must bear the cost of ther apy.It is there fore im por tant to dis cuss the cost of ther apy with the pa tient be fore dis charge.Risk of thera peu tic fail ure: Pa tient com pli ance with treatment while in the hos pi tal is taken for granted.How ever, once out side the hos pi tal this be comes much more dif fi cult to ensure.Poor com pli ance with the planned oral treat ment regimen could re sult in treat ment fail ure or re lapse of in fec tion.Re ad mis sion of the pa tient to the hos pi tal would quickly off set any cost sav ings re al ized by a change to oral ther apy.
An other po ten tial prob lem is that in com plete or in ade quate treat ment of in fec tion may con trib ute to the de vel op ment of mi cro bial re sis tance, mak ing it nec es sary to use more ex pensive or pos si bly more toxic agents to treat in fec tion.
In a study that spe cifi cally ex am ined pa tient com pli ance with an oral regi men, Paladino et al (21) docu mented an 81% com pli ance rate in out-patients tak ing cipro floxa cin twice daily fol low ing ini tial treat ment with in tra ve nous an ti bi ot ics in hos pital.Thera peu tic out comes in this group were ex cel lent despite com pli ance rates of less than 100%.Com pli ance rates were also ex am ined by Cra mer et al (22), who showed re sults iden ti cal to those de scribed by Paladino  dos ing, while slightly higher com pli ance rates were seen when medi ca tion was given once daily (87%).When dos age fre quency was in creased to four times a day, com pli ance rates dropped dra mati cally, ie, from 87% to 39%.Even if the pa tient is com pli ant, thera peu tic fail ures may still oc cur be cause of drug in ter ac tions that de crease the bioa vail abil ity of oral an ti bi ot ics.It is im por tant that all prescrip tion medi ca tions be re viewed with the pa tient at the time of dis charge from hos pi tal to en sure that no un to ward drug inter ac tions oc cur while the an ti bi ot ics are be ing taken at home.It is also im por tant to re view the use of any non pre scrip tion agents.For ex am ple, ant ac ids have been shown to re duce the ab sorp tion of oral qui nolones, tet ra cy cline, met roni da zole and cefpo dox ime proxetil; thus, pa tients should be ad vised to sepa rate ant acid use from in ges tion of these an ti bi ot ics by at least 2 h (23).The use of oral iron prepa ra tions should also be avoided in pa tients be ing treated with qui nolones and tet ra cyclines be cause iron de creases the ab sorp tion of these drugs (24).In ges tion of milk and other dairy prod ucts can also in terfere with the ab sorp tion of oral tet ra cy cline and, to a lesser extent, of qui nolones (25).

CRI TE RIA FOR SE LEC TION OF PA TIENTS AND DRUGS FOR SAT
Phy si cians have al ways been more com fort able us ing parenteral rather than oral drugs when treat ing se ri ous in fec tion, ow ing to con cerns about drug ab sorp tion, bioa vail abil ity, and se rum and tis sue lev els when oral agents are used.In the selec tion of pa tients for con ver sion to oral ther apy, sev eral cri teria should be ful filled: the pa tient should be he mo dy nami cally sta ble, able to in gest and swal low oral drugs and should have a function ing gas tro in tes ti nal tract.Yet, al though oral antimicrobial ac tiv ity can of ten be con sid ered as early as three days af ter ini tia tion of in tra ve nous ther apy in sta ble pa tients, some in di vidu als are not given oral ther apy un til they are discharged, ie, usu ally af ter at least seven days of in tra ve nous treat ment.
There are nu mer ous re ports in the lit era ture of the success ful use of in tra ve nous fol lowed by oral an ti bi ot ics for treat ment of se ri ous in fec tions.Many in volve in tra ve nous to oral se quenc ing within the same class of drugs, while some in volve a change in drug class.Among the con di tions treated with SAT are pneu mo nia (both com mu nity and hos pi tal acquired), pye lonephri tis, sep tic ar thri tis, os teo mye li tis, and skin and skin struc ture in fec tions.
If the above cri te ria are ful filled, the next step is to se lect the an ti mi cro bial that is most ap pro pri ate.The main thera peutic ob jec tive in chang ing from in tra ve nous to oral ther apy is to ob tain se rum and tis sue an ti mi cro bial ac tiv ity that is com pa rable with that ob tained with the in tra ve nous for mu la tion (26,27).Many oral agents have bioa vail abili ties that are similar to those of their par enteral forms.These in clude such drugs as met roni da zole, clin da my cin, chlo ram phenicol, flucona zole and cipro floxa cin (10,(28)(29)(30)(31)(32).Sev eral of these drugs have vir tu ally equiva lent bioa vail abil ity whether given in tra venously or orally.
Use of these drugs as part of a se quen tial ther apy regi men can help to al le vi ate the phy si ci an's con cerns about subop timal drug con cen tra tions when the switch to oral ther apy is made, be cause it is gen er ally ac cepted that these oral agents pro vide ade quate thera peu tic ef fi cacy (10,11,27).Some an tibi ot ics with oral bioa vail abili ties lower than those of their in trave nous for mu la tions have nev er the less been use ful in se quen tial ther apy.Such is the case for trimethoprim-sulfamethoxazole, ampicil lin and ce fu rox ime axetil (9,10,33).Much has been writ ten about the phar ma coki net ics of intra ve nous and oral cipro floxa cin sup port ing the use of the oral form in in tra ve nous to oral se quenc ing (30,31,34,35).For exam ple, with an av er age un im peded bioa vail abil ity of 75% (in the ab sence of sub stances that in ter fere with ab sorp tion), an oral dose of 500 mg cipro floxa cin pro vides an amount of drug equiva lent (ie, sta tis ti cally simi lar to the area un der the curve) to that ob tained with a 400 mg in tra ve nous dose (36).
Other qui nolones, such as ofloxa cin (not avail able parenter ally in Can ada), also have nearly iden ti cal phar ma cokinetic char ac ter is tics when given in tra ve nously and orally, and all qui nolones ap pear to have high vol umes of dis tri bu tion and rela tively low pro tein bind ing (31).Most qui nolones at tain tissue con cen tra tions that ex ceed the mini mal in hibi tory concen tra tion val ues of the com mon aero bic patho gens (37).
When se lect ing an oral agent, the phy si cian should take into ac count a number of fac tors in clud ing bio a vail abil ity, clinical ef fi cacy, tol er abil ity and cost.

EVI DENCE SUP PORT ING SAT
Rules of sci en tific evi dence that can be used to as sess pub lished data have been de vel oped and pub lished as part of a se ries of clini cal epi de mi ol ogy rounds (38).Six cri te ria are used to as sess ar ti cles and to de ter mine the va lid ity and appli ca bil ity of the re sults.These are: 1. Was the assignment of patients to treatment truly randomized?
2. Were all clinically relevant outcomes reported?
3. Were the study patients recognizably similar to your own?
4. Were both statistical and clinical significance considered?
5. Is the therapeutic manoeuvre feasible in your practice?
6. Were all patients who entered the study accounted for at its conclusion?
A lit era ture search re vealed 32 pub lished stud ies of in fections treated us ing SAT.Of these, 13 were ran dom ized controlled tri als and 19 were non ran dom ized.These stud ies are sum ma rized in Ta bles 2 and 3, re spec tively.
The rules of evi dence "con sti tute ap plied com mon sense and are de signed to maxi mize the ef fi ciency as well as the accu racy" of one's jour nal read ing (38).With this in mind, the arti cles were first strati fied into those that would be sub jected to the rules of evi dence and those that would not.Since the critical is sue is whether SAT would per form as well as stan dard ther apy, the stan dard, or con trol arm of a com para tive study, ide ally should con sist of in tra ve nous ther apy.If both the experi men tal and con trol arms use SAT, the is sue is con founded.Also, since the ran dom as sign ment of pa tients mini mizes much of the bias as so ci ated with non ran dom ized clini cal trials, use of a ran dom ized con trol de sign is es sen tial.Of the 32 pa pers, only six meet these cri te ria (Ta ble 4).The rest are either ran dom ized con trolled tri als, but with in ap pro pri ate control arms, or are non ran dom ized tri als.
For the pur poses of this pa per, clini cally rele vant outcomes were suc cess ver sus fail ure of clini cal re sponse.Success in cluded both cure and im prove ment.Bac te rio logi cal re sponse per se was not con sid ered as im por tant and was viewed as a sur ro gate marker.Sta tis ti cal sig nifi cance has no bear ing on whether the re sult is im por tant, but sim ply re fers to the like li hood that a par ticu lar re sult was ob tained by chance.Con sid era tion of sta tis ti cal is sues took into ac count whether sta tis ti cal tests were done and, if so, whether a dif fer ence was found that was as so ci ated with P<0.05 or, if not sta tis ti cally sig nifi cant, whether the sam ple size was large enough to rule out a type II er ror.
Clini cal sig nifi cance, on the other hand, does re late to the im por tance of a par ticu lar find ing.A dif fer ence in out comes be tween treated and con trol pa tients "be comes clini cally signifi cant when it leads to changes in clini cal be hav ior" (38).In the case of SAT ver sus con ven tional ther apy, ei ther a dif ference in fa vour of SAT or no dif fer ence be tween them is relevant since, in the lat ter in stance, the im pli ca tion is that by us ing in tra ve nous to oral switchover money is saved, there is im prove ment in the qual ity of life for the pa tient, or both.
Among these six pa pers, the main flaws are in the sta tis tical con sid era tions.From the point of view of ex peri men tal design, the best study was that of Kalager et al (39).The con trol arm con sisted of two drugs -both ef fec tive against most aero bic Gram-negative ba cil lary patho gens.The sam ple size was the larg est, and the sta tis ti cal analy sis was the most rigor ous.Only three of the six stud ies (40)(41)(42) ex am ined pa tients with only one type of in fec tion, while the other three (39,42,43) stud ied pa tients with a va ri ety of in fec tions, eg, lower res pi r a tory tract, uri nary tract, and skin and soft tis sue in fec tions.
In one trial (not in cluded in any of the ta bles) both compara tive and non com para tive study arms were used (45).The com para tive arm ran dom ized pa tients to ei ther ofloxa cin or a third-generation cepha lo sporin (cef tazidime or cef tri ax one) for treat ment of pneu mo nia, uri nary tract in fec tion, or skin and soft tis sue in fec tion.How ever, of the 22 pa tients ran dom ized to re ceive ofloxa cin, only eight were given se quen tial in tra venous to oral treat ment.It is not clear from the pa per whether a di rect com pari son was made be tween this small group of eight pa tients and those re ceiv ing in tra ve nous ther apy alone with a third-generation cepha lo sporin.The authors do, however, state that "none of the eight sub jects who were ran domized to ofloxa cin and who re ceived ini tial par enteral ther apy de te rio rated when switched to oral ofloxa cin ther apy."In the non com para tive study arm, pa tients were treated with oral ofloxa cin only.

IM PLE MEN TA TION OF SAT
In hos pi tals where SAT pro grams are par ticu larly suc cessful, they are usu ally de vel oped jointly by the in fec tious disease and phar macy de part ments work ing in con junc tion with the Phar macy and Thera peu tics Com mit tee.To fa cili tate the in tro duc tion of SAT into a par ticu lar hos pi tal, the ap pro pri ate in fra struc ture must first be cre ated.To do this, the col labo rative ef forts of mem bers of the phar macy, mi cro bi ol ogy, in fectious dis ease, nurs ing and ad mini stra tion de part ments are re quired.The mul ti dis ci plin ary per spec tive pro vided by these groups al lows the suc cess ful im ple men ta tion of rec om menda tions made by the Phar macy and Thera peu tics Com mit tee.
In te gral to this pro gram is the use of 's eque ntial ther apy remind ers'.These edu ca tional tools, de vel oped at the Van couver Hos pi tal and Health Sci ences Cen tre, are highly visi ble printed forms with the nec es sary in for ma tion on them (46).At the Hender son Gen eral Di vi sion of the Ham il ton Civic Hos pitals, for ex am ple, when a pa tient is started on a drug in tra venously for which an oral prepa ra tion is also avail able, the phar ma cist sends the se quen tial ther apy re minder, printed on bright yel low pa per, to the ward to gether with the in tra ve nous drug.The ward nurse then at taches this yel low sheet to the front of the pa ti ent's chart, where it re mains un til the in tra venous drug is ei ther dis con tin ued or changed to an oral form.An ex am ple of such a form is pro vided in Fig ure 1.In for ma tion rele vant to SAT is printed on the form, thereby pro vid ing an edu ca tional serv ice as well as a re minder to the phy si cian that oral ther apy should be con sid ered.Some of the chal lenges in the im ple men ta tion of SAT are re sis tance to change on the part of medi cal col leagues; a perceived in crease in work load by those in volved in man ag ing the SAT pro gram; and phy si cian re luc tance be cause of medical and/or le gal con cerns.Ex pe ri ence with SAT pro grams has shown that they are, in fact, rela tively easy to im plement and are read ily ac cepted by phy si cians and other health care per son nel pro vided that the ap pro pri ate legwork is done and the nec es sary in fra struc ture is first created.To help im plement SAT, it is im pera tive that col leagues be edu cated con cern ing an ti mi cro bial costs, sup port of chiefs of serv ices as well as col leagues be en listed, and con tinu ous sur veil lance and feed back to col leagues re gard ing the success of the pro gram be as sured.

CON CLU SIONS
Since the 1970s, medi cal lit era ture has docu mented the clini cal ef fi cacy of con vert ing pa tients from in tra ve nous an ti biotic ther apy to oral an ti bi otic ther apy as early as three days after ini tia tion of in tra ve nous ther apy in sta ble pa tients.Yet, such pa tients tra di tion ally have not been given oral ther apy un til hos pi tal dis charge, ie, usu ally af ter at least seven days of in tra ve nous treat ment.The use of drugs that have vir tu ally equiva lent bioa vail abil ity in in tra ve nous and oral forms, such as met roni da zole, clin da my cin, flu cona zole and cipro floxa cin, can help al le vi ate phy si cian con cerns about subop ti mal drug con cen tra tions when con vert ing to oral ther apy, and can perhaps in crease the ac cep tance of SAT.A criti cal re view of the medi cal lit era ture sup ports the role of SAT.What is also clear, how ever, is that fur ther stud ies are nec es sary to de ter mine the ideal time for in tra ve nous to oral con ver sion and fac tors that may limit or im pede the use of oral ther apy.
We con clude that SAT is not only an im por tant tool for re aliz ing sub stan tial cost sav ings in the treat ment of pa tients with se ri ous in fec tions, but can greatly add to pa tient com fort and

TA BLE 1 Ac qui si tion cost of one dose of com monly pre scribed an - ti bi ot ics* Drug Dose (route) Ac qui si tion cost ($)
et al for twice-daily Prices sup plied by pur chas ing de part ment, Hender son Gen eral Di vi sion, Ham il ton Civic Hos pi tals, Ham il ton, On tario, De cem ber 1993.IV In tra ve nous; PO Oral *MAN DELL et al 308 CAN J IN FECT DIS VOL 6 NO 6 NO VEM BER/DE CEM BER 1995

TA BLE 4 Clini cal tri als com ply ing with rules of sci en tific evi dence Author (ref er ence) Ran dom ized con trolled trial Study pa tients simi lar to your own Con sid era tion of is sues Ma noeu vre fea si ble in your prac tice All pa tients ac counted for Sta tis ti cal Clini cal
tiv ity.Hos pi tals must play a pro ac tive edu ca tional role in im ple ment ing SAT pro grams if this im por tant thera peu tic strat egy is to be come the fu ture stan dard of care.Prepa ra tion of this pa per was sup ported by Bayer Inc Health care Di vi sion.
Fig ure 1) An ex am ple of se quen tial ther apy re mind ers (ref er ence 46) CAN J IN FECT DIS VOL 6 NO 6 NO VEM BER/DE CEM BER 1995 pro duc AC KNOW LEDGE MENTS: