The Authority/Pharmacotherapy Care model: an explanatory model of the drug use process in primary care⋆

1. Introduction 

Many models have been developed to describe various aspects of drug utilization, the first of which were those of Brodie1, 2 in the United States and Crooks3 in Britain. Few of these models have addressed the drug use process in a comprehensive fashion. Most often, their focus has been on a single aspect of use (eg, appropriateness as judged by criteria)4, 5 or have adopted from a single perspective (eg, 1 antibiotic in 1 hospital, as viewed by the pharmacy and therapeutics committee)6, 7. Often, outcomes or consequences of use are ignored. However, drugs—and drug use—are part of a dynamic system. Therefore, to truly understand the use of drugs, we need to examine all of these factors together and not artificially separate or isolate them.

Those studies that examine drug consumption on a population or national level tend to be quantitative, focusing on numbers of prescriptions or medication units dispensed while ignoring other scenarios involving drug use, such as prescribing.8, 9, 10 Some, such as drug utilization review,1, 2, 5, 6, 7 focus on qualitative aspects of drug use (ie, appropriateness) within an institution while little attention is paid to the locus of inappropriate drug use or the responsible factors.

All of the above studies have been referred to as “drug utilization studies,” “drug use studies,” or “drug usage studies” or even medication use evaluations.8 There is no real consensus on terminology; all of these terms are used interchangeably. Currently, no consistent taxonomy or model exists that accurately and completely categorizes different types of drug studies. Thus, there is confusion over what is meant when people refer to a drug utilization study. In addition, the focus is often at the population level; few address the situation of the individual in a primary care situation.

A need therefore exists for a model that incorporates all of the steps of the drug use process and accounts for inputs from different influences on drug utilization. As well, such a model should explain relationships between factors and where they fit into the overall scheme. It should account for drug use in terms of prescribing, dispensing, consumption, as well as drug marketing and distribution. Finally, it should have direct application to the primary care of patients.

A structural-functional model is presented that incorporates all of the above factors into a single, comprehensive, and representative matrix. Its purpose is to depict the processes and relationships that exist in the drug use process within primary care. It is a model that depicts the structure of relationships between the various players in the drug use process, hence is structural. It describes the functions of each level of care, and the outcomes that these functions produce, hence is functional. Finally, it describes the flow of events over time. The perspective is that of the individual who experiences a health-related problem requiring drug therapy, and proceeds through the primary care system.

2. The Authority/Pharmacotherapy Care model 

Figure 1 depicts the Authority/Pharmacotherapy Care (APC) model. In this model, we are concerned with the people and processes involved with drug use at the level of the individual in a primary care situation. Thus, the term “pharmacotherapy” has been used, because we are interested in all aspects of care revolving around drugs, and it therefore includes all people involved in that process. Thus, the perspective is that of the individual patient who interacts with an individual clinician, such as the one involved in the delivery of pharmaceutical care to specific patients. It should be noted that the term “pharmaceutical care” is used to denote drug-related care provided by the pharmacist to the specific patient, and for which the pharmacist is responsible. It forms part of the overall pharmacotherapeutic process. The latter term is all-encompassing, incorporating not only pharmaceutical care but also the contributions and responsibilities of the physician and patient.

View Large Image Download to PowerPoint

Figure 1. Model of the drug utilization process for an individual (Authority/Pharmacotherapy Care model). Boxes at each level are divided into 3 categories, which are (reading left to right) structures, processes, and outcomes.

The model is eclectic, incorporating individual levels of care and, for its analytical focus, Donabedian's evaluation model for health care services.11 It was named to reflect its underlying premise in that it describes the structures and relationships involved in the drug use process in terms of authority that is transferred from the patient, to the physician, to the pharmacist, and then back to the (enlightened) patient. The relationships within the model depend upon the transfer of authority from one person to another and involve the use of drugs in patient care, hence the acronym APC.

The APC model is also a functional model that describes the basic activities that are involved in the drug use process, thereby identifying the points in the health care process where quantification and evaluation of drug use take place. These points in the process are also points of potential influence. They indicate the loci where we derive pharmacoepidemiologic data and where we can apply interventions aimed at improving the whole process.

The APC model is founded on 3 key concepts: (1) authority, (2) transfer of authority, and (3) responsibility of patients and other members of the team. All of these concepts are essential to the drug use process. Each will be discussed in detail below.

3. The concept of authority 

The current atmosphere in western society is one of personal control, responsibility, and authority over one's own self. It is analogous to the ethical concept of autonomy. Thus, there is an intrinsic authority in that no person may even touch another person without the consent of that person. Individuals are free to care for themselves (within the law) and, in normal circumstances, only enter the health care system when they choose to because they have the ultimate authority over themselves. Exceptions include individuals who require health care but are not able to do so for themselves, such as those who are seriously hurt, unconscious, too young (ie, children), who are mentally incompetent, or who pose a risk to themselves or others. In these cases, a surrogate is required to assume responsibility and act on the other person's behalf.

A second form of authority is vested in the health care practitioner through education, skills, and licensure. Society permits persons to practice as professionals in a given jurisdiction, which, in the model presented here, applies to physicians who have the authority to diagnose and prescribe (including medical specialists) and pharmacists who have the authority to dispense medications. Other prescribers, such as pharmacist practitioners, midwives, or nurse practitioners, could be included in the model in the place of the physician, depending on the nature of the situation. However, these professionals cannot perform their services unless the individual having the authority allows them to do so by transferring that authority to them. That is, in the usual case, the patient must initiate the transfer process.

4. Transfer of authority 

The individual remains responsible for her/his own welfare, but may transfer part or all of this authority to those who have requisite qualifications to practice as health care professionals. When the individual perceives symptoms, she/he acknowledges that there is a medical problem, then accepts that something must be done about it. There may be an attempt to self-treat or to seek help outside of the traditional medical system. However, the present model assumes that, when such attempts end in failure, the individual recognizes the necessity of entering the health care system for help. That is the process of becoming a patient.12, 13 It requires that such persons accept the patient role in which they transfer authority for treating the perceived illness to the medical practitioner. That process is accomplished by consulting a physician. Other possibilities include being brought to a hospital by ambulance, police, or court order, but the process is essentially the same in that authority for treatment must be appropriately obtained. The transfer of authority continues through the model, as is detailed below in the following sections.

It should be noted that the transfer is not absolute and not complete. The person transfers authority to allow the next phase to begin. As in ethical consent for research, it can be revoked at any time. Similarly, the person receiving the authority cannot exceed the limits of that authority without seeking permission. Thus, the individual maintains the ultimate authority.

5. Responsibilities of patients 

When a person transfers authority to a practitioner, she/he allows that practitioner to exercise privileges granted according to law and custom.12, 13 That means allowing the practitioner to perform professional activities such as, in the case of a physician, a physical examination, laboratory tests, or surgery. Consent forms (eg, for surgery) represent the embodiment of the transfer of authority. In the case of a pharmacist, it would involve providing personal-health–related information that would be necessary in determining the appropriateness of the therapy or that might draw attention to a potential drug-related problem. In any case, practitioners can only proceed with patient care when they have received the requisite authorization, either explicitly or implicitly.

When the role of patient has been assumed, the individual has the responsibility to follow the advice of the practitioner and to provide information truthfully. That includes adherence to treatment plans including the taking of drugs.12, 13, 14 The individual also has the authority to question or disagree with the advice, but has the responsibility to raise these issues with the health care provider.

6. Structure of the model 

The APC model integrates the concepts of authority, transfer of authority, and the importance of the patient's role in drug use. It is explanatory in nature and has 5 levels of authority and functioning, corresponding to 5 sequential levels of patient care. These care levels and their relationships are presented in Table 1. The first care level is that of the individual, who must perceive an illness, then consult a practitioner for help. The second care level is that of the physician, the third is that of the pharmacist, the fourth is that of the patient, and the final level is that of the drug.

Table 1. Categorizing drug use studies using the APC model

	Care Level	Subject	Structure	Process	Outcome
	1	Individual	Epidemiology of numbers of patients or diseases	Knowledge, attitude, behavior of members of the public	Number or rate of patient consults
	2	MD	Numbers of physicians, specialists, categorized by sex, age, geography, etc.	Quality of prescribing (appropriateness); quality of interaction	Number or rate of prescriptions written
	3	Pharmacist	Numbers of pharmacists, by sex, age, geography, etc.	Dispensing errors; appropriateness of information provided	Number or rate of prescriptions filled
	4	Patient	Number of prescriptions received; number of potential drug consumers	Appropriateness of consumption or adherence	Rates of actual consumption or adherence
	5	Drug	Types or quantities of drugs used (drug use surveys); patterns of use, comparisons	n/a	Rates of effectiveness, ADRs; cost-effectiveness

ADR=adverse drug reaction; MD=medical doctor; n/a=not applicable for drug utilization studies because it deals with pharmacology within individuals; however, it is a necessary step in the drug use process in that it determines ultimate patient outcomes.

The subdivision of each level of care has been based on the approach to evaluation developed by Avedis Donabedian.11 He determined that, to evaluate a model or a professional service, one must address 3 of its aspects, which are structure, process, and outcome. Each care level in the present model is divided into 3 contiguous categories reflecting those 3 components, ie, structures, processes, and outcomes of care at that particular level.

6.2. Processes 

Processes are the activities undertaken by the structures in response to the receipt of authority. In Care Level 1, the activities of the individual are to (a) perceive that an illness exists, (b) acknowledge that the illness requires the input of a physician, and (c) accept the role of patient. Accepting that role terminates in consulting a physician, and marks the transition from a disengaged and independent individual to a co-dependent patient who agrees to submit to medical procedures and follow advice provided by the new authority figure.12, 13 Those latter activities constitute the outcomes for Care Level 1, which is further discussed below.

Processes are acted upon by influencing agents, labeled I in Figure 1 and are further elaborated upon in Figure 2. Persons take action based upon both their background (ie, personal qualities) and factors, which are specific to the situation, that is, those factors are linked to the action of consulting the physician or taking the drug. Both are influenced by agents acting as “push forces,” such as advertisements, which are urging or promoting the taking of drugs. Other agents, such as laws or regulations, act as inhibitors.

View Large Image Download to PowerPoint

Figure 2. Agents and factors influencing the drug utilization process.

At each level of the model, all of those influences impact upon the processes involved in drug utilization, interacting with each of the individual caregivers in different ways. They may serve to increase demand for drugs in general or for specific types or brands of drugs, which may or may not be appropriate. The Influences (I) have subscripts 1 to 5, corresponding to the respective Care Levels, indicating that there are different influences affecting each process. I1 represents the influence on individuals, as they become patients within Care Level 1, while I2 and I3 are those acting upon the physician and pharmacist in Care Levels 2 and 3, respectively. I4 is a second influence upon the patient, but under new circumstances created through interaction with the health care system and the assumption of the role of “patient as care manager” (ie, Care Level 4). I5 is the final influence, which depends on the pharmacokinetics and pharmacodynamics of the drug as well as the physico-chemical properties of the person, which may affect drug action, such as body mass, proportion of fat, metabolic rate, organ function, concomitant drugs, genotypic and phenotypic characteristics, etc.

7. Study classification using the APC model 

The APC model can be used to classify drug utilization studies. Because of its explicit layout, it identifies the locus of a drug use study and its influence. Therefore, it can be used to categorize these studies so that they can be better understood. For example, there are studies that evaluate Level 1 of care to determine how effectively patients acknowledge the need to seek professional advice for medical problems that require drug therapy. In Care Level 2, we can identify studies that focus on how appropriately physicians prescribe medications. In Care Level 3, we can determine how effectively a pharmacist provides the correct medications and appropriate information to patients. In that way, the model provides a common understanding of the different types of studies that examine the use of drugs in society.

Table 1 identifies aspects of drug utilization that may be assessed by various types of studies, and categorizes them by level. For example, the focus of Care Level 1 evaluations is on the patient and patient demands for medical care and for prescriptions, as well as on the appropriateness and economic impact of such demands.

By identifying focal points where drug utilization may be assessed, the model can guide in screening for potential problems. Thus, it can be a useful tool for a clinician, researcher, or manager in problem identification. The investigator can then assemble evidence and pinpoint the precise location of successes and failures within a practice, or for specific group of medications, and more precisely identify where attention needs to be paid.

Drug utilization studies are conducted for a number of reasons, hence a variety of models have been used. They can be either quantitative, where they summarize drug use, or they may be qualitative, assessing that use against criteria. Studies using social science qualitative research–related methodologies can also be included within the broad label of drug utilization studies. The focus could involve a single drug, prescriber, or patient, or a comparison between 2 or more. In fact, combinations of those aspects may be dealt with in any given study, depending on its purpose. In addition, there may be correlational studies that examine factors related to drug use or to inappropriateness. Finally, some studies represent interventions that attempt to optimize or improve utilization. Utilization may be evaluated at each segment (ie, structure, process, and outcome) of each level throughout the entire model.

Health surveys or epidemiologic reports of specific diseases (eg, depression)17 could be considered as Care Level 1 analyses, focused on structures. They provide an estimate of the total potential number of patients in the system. With respect to processes, Lawton18 recently reviewed papers describing the experiences of the public with respect to their health and illnesses. Utilization studies conducted in emergency departments have examined the appropriateness of drug use and the outcomes from that utilization.19 Factors related to utilization of physician practices were examined by Wagner and Hendrich,20 who found that all patients rated as misusers to any degree represented only 17.1% of the population, yet accounted for 42.7% of the visits. Others examining influencing factors include Andrade et al,21 who reported that the change in status of H-2 blockers did not affect physician visits. Similarly, studies of access to care22 would fall into this category.

Care Level 2 focuses on physicians and their activities, which include medical care and prescribing. Studies of physician numbers, such as that by Anderson and Rosenberg23 are examples of analyses of structures. Processes have been the focus of many drug utilization reviews and prescription audits.24, 25, 26 Outcomes have been addressed in research such as that by Melander et al.27 Investigations into the influences on prescribing have included clinical,28 personal,29 situational,30 and sociological factors,31 among others.

Care Level 3 is concerned with pharmacists and their activities including dispensing and pharmaceutical care. The presentation of models of pharmaceutical care (PC), such as those of Bajcar et al14 and by Sorensen and Biebighauser32 would be categorized as structural. Pilnick33 has done a process evaluation in his research of approaches used by pharmacists to deliver pharmaceutical care. Pharmaceutical care outcomes in elderly patients have been examined by Sturgess et al.34 Studies dealing with access to drugs (eg, willingness to pay35 or ability to pay/access36) or change of reimbursement status, such as formulary delisting or Rx-OTC switch,37 would be classified as I3 studies. These are all examples of external agents influencing the supply of pharmaceuticals.

Care Level 4 involves the patient and patient's activities related to the taking of drugs such as adherence to the prescribed regimen and actual drug consumption. For example, Venturini et al38 studied that outcome in diabetics and Iskedjian et al39 did so with hypertensives. Influencing factors were examined by Iskedjian et al,39 Kuzujanakis et al40 and Shalansky and Levy.41

Care Level 5 focuses on the effects produced by drugs, including both beneficial and adverse clinical effects of drugs and drug interactions.42 Although not often considered as drug outcomes, pharmacoeconomics can also be included in Care Level 5.43

However, evaluations can have multiple foci and some studies may be quite difficult to categorize because they may address several different topics in one paper. Examples are the evaluation of a service (eg, pharmaceutical care)14 or the study of some aspect of the drug use process such as examining the relationship between outcomes (eg, adverse drug reactions) and patient or prescriber characteristics.26 For example, Hazra et al44 assessed the drug use process at Care Level 2 (ie, prescribing) and Care Level 3 (ie, dispensing). In such cases, all foci should be identified, and the study classified accordingly (ie, multifocused).

8. Evaluating drug use 

Evaluation of structural features of a model such as APC normally involves quantifying drug use in a given practice area over a given time period. It commonly involves a descriptive categorization of drug use patterns which Rosman and Sawyer8 have termed Drug Use Surveys. In other words, evaluation of structures and outcomes tends to be quantitative, rather than qualitative. They produce results that are numerical (ie, descriptive statistics), which may be useful for comparing between practices or for monitoring over time. However, they often tell us little about the appropriateness of those structures or outcomes.

Process evaluations often involve assessment of the quality of those processes, judged against a set of criteria of appropriateness. Examples are Drug Utilization Reviews4, 5 and Prescription Audits.3 Numerous models with practical examples have appeared in the literature.45, 46, 47, 48 Currently, preferred approaches are prospective and patient focused.

Outcome evaluations may be like structure evaluations (ie, descriptive quantifications) or they may involve comparisons between drugs or groups of drugs or against predetermined criteria for appropriateness. Petitta et al49 and Bunting50 have noted that outcome evaluations have become an integral part of the quality assurance program. Efforts to compare observed drug use outcomes with norms have been implemented. One example is the Prescription Analysis and CosT reports program in Britain, where individual prescriber outcomes are compared “with those of peers as well as with local and national statistics.”51 It, however, has been criticized because it uses only quantitative results (especially costs) and ignores qualitative aspects and clinical outcomes.

A benefit of the APC model becomes apparent here. If we focus only on outcome evaluation and not the structure or process and we find the outcomes are either positive or negative, we do not know why this is so. Perhaps the outcomes are related to the structures or the processes being good or bad. Also, if we consider the existence of the multiple integrated levels of primary care and find that the outcome of care is not as expected, we can identify the level of care at which the gap occurred (ie, patient Care Level 1, the prescriber Care Level 2, the pharmacist Care Level 3, or the medication use Care Level 4).

Another way to view the transition between levels is to regard outcomes from one level as the workload statistics for next level. Thus, the output from Care Level 1 (ie, patients) becomes the workload for Care Level 2 (the physician). The output from Care Level 2 is written prescriptions, which form the workload for the pharmacist in Care Level 3. The output then is prescription drugs, which are to be consumed by the patient in Care Level 4.

Discrepancies occur when there are dropouts as people progress through the model. In Care Level 1, dropouts are people who identify an illness, but fail to consult a physician. Sometimes, the patient recovers uneventfully, in which case the number of physician consults will be fewer than the number of cases of perceived illnesses. In other situations, the patient becomes more ill and is forced to enter the system as a patient. The patient can fail to take the prescription to the pharmacist, pick up the drug, or take the drug. Therefore, there will be differences between numbers of drugs prescribed, dispensed, and consumed. This discrepancy is a pervasive problem in drug utilization studies, and sometimes in randomized controlled trials, which remains largely unresolved. That is, we are never certain exactly how many drugs are actually taken. The APC model helps to identify this potential discrepancy and highlight the importance of integrated approach to DUR.

Evaluation of influencing agents (identified as I in Figure 1) involves the examination of the effects of any of those on the drug use process in any of the process domains for each level of care. As with the other aspects of this model, one could examine the structures. For example, one could determine the types of advertisements that were used to promote a given drug or class of drugs, or costs of such advertising. One could examine the processes against a set of criteria (eg, determine whether such advertising was ethical) or the outcomes (ie, did the advertising increase consumption of that drug?). Similarly, one could examine any of the influencing agents.

Personal qualities play a role in drug utilization, as do specific factors. They can determine a person's motivation for taking drugs and for adherence (or nonadherence) to prescribed regimens. Such information may be of interest to medical sociologists as well as to marketing departments of pharmaceutical firms. As with other elements within this model, one can examine their structures (eg, knowledge base, attitudes, beliefs), processes (eg, usual activities, perceptions), or outcomes.

However, in any evaluation, it is important that the impact of interventions be measured at a point adjacent to the location of that intervention. In fact, an error often made when assessing a service or program is selecting an inappropriate aspect of care. Evaluation must be based on an indicator that is proximal to the activities for which the individual being assessed is responsible. Thus, a pharmacist engaged in pharmaceutical care can be seen to be operating in Care Level 3. Therefore, if pharmaceutical care is to be assessed, the outcome being evaluated must be a Care Level 3 outcome. In the past, some evaluations of pharmaceutical care have focused on patient outcomes that have been far removed from what was actually done by the pharmacist. As a result, no impact was observed in many analyses,52, 53, 54 which used outcomes from Care Level 5, rather than Care Level 3. Therefore, it is important to refer to the model to determine the appropriate level and locus for interventions or for data collection.

9. Using the APC model 

Table 2 presents a 5-step approach to using the APC model, along with an example. Suppose the manager wished to evaluate an office-based pharmacist caring for elderly patients using a patient-oriented pharmaceutical care model of practice. Step 1 would be to identify the object of the evaluation, which, in the present case, is the pharmacist. Step 2 is to classify the level and function of the evaluation. In the present case, we are interested in evaluating the pharmacist, which is Care Level 3, and delivery of care represents a process. In Step 3, we identify specific aspects to evaluate. In this case, we would like to examine the quality of the drug regimen in the elderly patient (who requires an antidepressant), for which the pharmaceutical care pharmacist is responsible. In Step 4, we establish objective evaluation criteria against which performance can be measured. We would refer to official guidelines for drug therapy in this population, such as those of the Faculty of Old Age Psychiatry Working Group of the Royal College of Psychiatrists.55 The final steps would be to select a sample of patients, identify the drug regimen, compare it with the guidelines, and note discrepancies.

Table 2. Using the APC model to evaluate utilization against criteria

	Step	Activity	Example
	1	Identify object of evaluation	Impact of a pharmacist on prescribing
	2	Classify its level and function (structure, process, or outcome)	Care Level 3 process
	3	Identify specific aspects to evaluate	Quality of drug regimen, especially antidepressants, in an elderly patient
	4	Determine evaluation criteria	Guidelines of psychiatric association
	5	Collect data, analyze, feedback	n/a

By using the model to locate the level of the evaluation, the evaluator is able to immediately identify the elements of concern. As well, the items used in the assessment can be identified and the evaluator can assure that they are appropriate (ie, at the same level).

10. Limitations 

There are limitations to using this model. It is limited to the microperspective, where events are based on those of the individual. It focuses on the patient as he/she progresses through the system. It does not deal with more global points of view, such as policy, formulary, or guideline development, which have a macrofocus. There must be adaptation for use for other purposes.

11. Conclusions 

A structural-functional model of the drug use process has been presented. It represents a straightforward depiction of the traditional process of drug taking but can be adapted for other scenarios. Most of the points of focus for drug utilization studies are represented in this model; consequently, it may be used to develop an understanding of the factor involved in the process and the relationships between factors.

The APC model provides a frame to help orient and coordinate professionals who are studying medication use at different levels and to facilitate the transfer of knowledge about the structures, processes, and outcomes of one level to the individuals who may be studying the subsequent level. The model also provides a common framework that could facilitate dialogue among different disciplines and foster more interdisciplinary research to address issues in medication use in our society. It clearly describes the notion of different levels of authority and accountability for each level of care. It illustrates that care needs to be taken in designing intervention studies to maintain a proper focus. As well, one must take care when selecting outcomes and measurements to assess these interventions to avoid selecting an indicator that is too far away from the level of care that the intervention is meant to influence. Clinicians and researchers alike can benefit from the APC model.