Thursday, October 17, 2019

A Practical Quality Agenda in a Primary Care Practice

If you ran a primary care practice, how would you approach the issue of quality of care?  I faced this issue in my practice, and I thought about it in depth.  Twelve years ago I wrote this paper on how I thought about it.  It's very practical.  Much of what you read here you will not have read elsewhere.  I commend it to your attention.

One note - although much of it holds up despite the time lapse between when I wrote it and now, EHR's were not widespread then; now, they are ubiquitous.  So that part of the paper is partially outdated.  But other than that, as I read it over, I think much of it still holds true. 

If you want to improve your practice quality, you will need a roadmap.  I think this paper might help.

Editorial note: I have 48 footnotes for this paper, but somehow it's hard to put them into this blog, but trust me, what I say is well founded!



We have known for some time that our medical care system is failing to implement what we know and what is theoretically possible.  The Institute of Medicine found that the medical enterprise has outgrown the capacities of our basic institutions, the great variability of care implies quality deficiencies, indicated preventive measures are inconsistently applied, diffusion of knowledge is chillingly slow, and indicated care is offered a random adult patient at an estimated rate of 55%.  Most recently, the UCLA-Rand group has extended their findings to indict ambulatory pediatric care as well.  They conclude: “Strategies to reduce these apparent deficits are needed.”

Indeed they are.  Our primary care group (mostly pediatrics) has been addressing ourselves to quality improvement (QI – used here to denote general quality efforts rather than any specific method) for five years in an organized way, and in doing so have discovered approaches that might help others to find paths for QI for their practices.

Our first discovery was that there is indeed no current roadmap; we could find no articles in the literature examining QI in offices in detail.  If only we had the know-how to diagnose and treat practice as we do a patient: take a history, do a physical examination, take measurements, and proceed with well established plans for prevention and treatment!  But quality science has not yet provided office practices with the equivalent of even vital signs, let alone a diagnostic outline, a Practice Quality Index, or established modes of action.  Consequently, if you are designated the Quality Officer in a group practice, at the moment of your appointment, you face near-virgin territory, and will need to invent and discover your own path forward.

Practice roadmaps will differ, of course, since practices differ widely in their abilities, needs, proclivities, visions, patient populations, and organizational structures.  Nonetheless, every practice seeking an organized approach to QI will most likely need to ask similar questions.  For the question, “What area does quality encompass?” we have found three separate sectors.  For the question, “What modalities of action do we have?” we have found two distinct modes.  For the question, “What are the constraints on our activities?” we have found three impedimental areas.  And for the question, “What practical projects suggest themselves?” we have found so many that there will clearly be no end to this quality journey.


Pay For Performance (P4P)

Pay For Performance (P4P) is the most conspicuous current embodiment of QI.  P4P identifies elements of quality that are important and which are putatively easily measurable through claims data, thus finessing the difficulty of actually measuring quality on a non-research basis.  P4P then rewards either health plans or individual practices that score well on the measures chosen.  A practice will be tempted simply to take the P4P measures, aim at them, and call that their program.

To do so, however, might be a mistake.  The elements of practice are generally important, true, and the financial rewards are of course attractive.  But P4P has marked weaknesses.  P4P measurement validity is itself questionable.  Scoring well on P4P items does not indicate general practice quality, especially if the practice “teaches to the test” of P4P indicators, and concentrating on P4P may well “crowd out” more important QI efforts.    “P4P programs are being implemented in a near-scientific vacuum.  There is scant evidence about the appropriate focus, effectiveness, and the general circumstances in which they may work best.” Thus, although P4P might well turn out to be an element of a practice’s QI program, it will be too thin a reed upon which to build an entire QI program.

Three Realms of Quality

A wider view of practice quality can be obtained from looking closely at the definitions of quality.  Most obvious to us is biomedical quality, defined by the Institute of Medicine: The degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge.   A second area, the way in which care is provided, is also defined by the IOM: timeliness, patient-centeredness, safety, effectiveness, efficiency, and equitableness.  These qualities would apply both to how clinicians function (tending to labs and X-rays promptly, involving patients in decisions, etc.) and how the office as a whole functions (waiting times, respect from staff, ease of navigating the system, etc.).

The third definition of quality would be even wider: How well a practice helps its patients.  Unlike definitions one and two, this wider definition includes the humane services of counseling, befriending, understanding, and supporting patient, not only directly in the office, but by referral to allied resources. 

As scientific clinicians we would adopt objectives from the first definition and the clinical aspects of the second; as directors of a practice we need to consider the second part of the second definition, the office experience for patients; and as humanitarian clinicians we want to consider the third definition.  Clearly, then, when faced with the possibility of choosing objectives for QI, a practice can choose from a very wide field.

Two Modes of Intervention

Each QI objective chosen will need a matching means of intervention.  The plethora of approaches to QI  can be divided into two types.  Professional Enhancement (PE) aims to improve the mind, skill, and motivation of the clinician.  Systematic Reengineering (SR) aims to improve the operating mechanism of the practice within which the clinician performs.

Clinicians almost reflexively tend to think of QI in terms of PE.  They want to improve their knowledge, memory, understanding, and motivation.   They think of following the literature, continuing medical education, certification and recertification, clinical performance assessment, being reflective, listening to and caring for each patient, searching for the hidden meaning behind patient visits, and detecting disease early from small clues.  Clinicians tend to remember instances when they stopped short in leaving the exam room when a small remark of a patient hit home, how they returned to the patient for more history and examination, and how two weeks later an operation saved a life.  They remember when a bit of knowledge popped up in their minds and led to a smart and important diagnosis.  They recall the meaning they bring to some patients lives, and vice-versa.  This is traditional, palpable quality.

Health services researchers and administrators, by contrast, tend to think in terms of SR.  SR comes from the industrial quality experience, where establishing regular procedures aims at decreasing variability, with the mantra “do-it-once-do-it-right.”  The SR approach finds recurrent stereotypic situations within medicine’s endless complexity that can be choreographed for standard execution with low variation, and hence higher quality.  SR thinks of care plans, flow sheets, and forms.  The goal is to imbed the clinical team in a system that makes it easy, rather than a constant challenge, to do the right thing at the right time.  For system engineers, the clinician is sometimes thought of as a standard commodity. Instead of individuals, SR thinks of teams: “Ranking (of personnel) is a farce.  Apparent performance is mostly attributable to the system that the individual works in, not to the individual himself.”

A practice can and should use both approaches.  True, the effect of PE is more assumed than documented, in contrast to the well-measured impact of SR.   True, that the obstacles to effective education are remarkable.    Nonetheless, some of the assumed superiority of SR must emanate from the more easily measurable nature of SR applications.  Stereotypic situations are inherently more easily measured than the qualities PE aims at -- human relations, complex diagnoses, unique situations, and judgments as members of adaptive organizations.  Sometimes SR will be appropriate -- narcotics control is better handled by a computerized check-in, check-out system than by education on the harms of addiction.  But other objectives require PE --- education, training, and reinforcement are more appropriate to improve empathy than signs on the wall saying “We Care.”  Both matter: the quality of the clinician, and the system in which he or she operates.

Three Constraints, and Possible Solutions

If there were not severe constraints on the abilities of practices to pursue a QI program, programs would be popping up everywhere, yet even in academic faculty practices are “in an infancy stage” of QI.  The constraints on QI actions can be summed up as organization, cost and effort, and measurement.


Realistically, for a QI effort above minimal scale, it will be easier to envision an organized QI effort than to actually accomplish it, because highly organized group practices are a rarity.  Most small groups (groups of six or fewer physicians deliver about 2/3 of the primary care in the United States) are partnerships with individualistic physician-partners, which are inherently hard to lead.  While many physicians are natural leaders, few are trained in management and quality science.  Moreover, the QI effort will require time and energy to analyze needs, to meet and communicate, to strategize, to revise practice protocols, to produce practice aids, to track progress, etc.  The time taken will need to be stolen either from clinical time (which, unlike QI work, is directly remunerative), or from scarce time off.  Yet most practices will not want to pay a clinical leader for this time and effort because of the traditional debilitating conceit of practitioners that only patient care should be compensated, and that administration is “pushing paper.” Some clinicians, fired by idealism, will enroll themselves as a “champion” of a QI project once, or perhaps twice.  But beyond that, without payment, the champion’s QI leadership will recede.

By the same token, QI innovation in office operations requires a sophisticated practice manager, one able to envision programs, collect and analyze data, and manage personnel with changing roles.  But such managers are scarce, and hospitals, specialists, and multispecialty practices compete strongly for their services.

To be well enough organized to pursue a serious QI program, the practice will need to make a leap of faith, and break away from prejudice.  It will need to support talent that is already in place, or recruit new organizational talent.  It will need to believe in the ultimate advantage of investing in organization.  The increased cost of an improved organization may not be directly recompensed by QI efforts themselves.  But as a side benefit, a highly organized and well-run practice may bring other advantages to the practice.  Better-run practices can become more productive and thus make more money.  Better-run practices can use mid-levels or contract physicians, and may expand to additional offices and extended hours.  They might expand into other allied clinical fields, such as travel medicine, or lactation classes.  They can negotiate with insurance companies in a more organized and successful fashion.  They may attract more talented graduating physicians.  Morale in well-organized practices can be superior.  They can become, in fact, the cream of the crop and face the future with increased confidence.  In other words, the solution to the organizational obstacle may be to face the music and invest in oneself.


Organizations run two kinds of costs.  Routine QI costs are those of maintaining competence, such as fulfilling a required 25 hours of CME per year, or reviewing employee performance in patient relations.  Investment QI costs involve more substantial efforts to upgrade and improve.  For example, the American Academy of Pediatrics has apparently endorsed 162 separate points of advice that a pediatrician should impart to a family over the course of pediatric care.  If a practice endeavored actually to follow the Academy’s directives, this Herculean effort would clearly be a significant effort to upgrade and improve its service.  The costs would occur both at the initiation of the project – change – and the ongoing effort – maintenance. 

From a business point of view, one would expect routine maintenance to be included in general overhead as “the cost of doing business.”  Investment costs, however, would need to be covered by enhanced revenues.  It could be that psychic “idealistic revenue” could cover the cost off one or two practice innovations, but over time idealistic energy would usually wane, and financial enhanced revenue would be necessary.  This would be equally true for for-profit and not-for-profit enterprises.

What makes QI investment particularly hazardous is the well-known difference between innovation in medical care and other fields.  In a classical free market, consumers recognize higher quality and may pay more for it, and/or higher quality at the same price accrues more business and increased profit.  Higher quality clinical care, however, is often invisible to patients, payors, and even colleagues.  Raising prices often accrues little increased revenue since prices are set by insurance companies.  If additional patients are attracted, many practices see little or no benefit, since they are often already very busy, and adding additional clinicians can yield little profit under the partnership model, since new clinicians quickly become partners and share in the proceeds, leaving little if any additional profit for the innovators.  Thus, investing in QI is often likely to reward patients with improved care and clinicians with psychic well-being, but to penalize practices with unrecoupable investment costs.

Solutions to the financial problem are at least threefold: (1) adopt changes that are low-cost; (2) discover changes that are billable and thus support themselves; and (3) find innovations that produce a joint product – for instance, a change that simultaneously improves quality and efficiency, thus enabling a practice to make more money indirectly.   Examples for solution (1) would be many PE interventions, such as journal clubs and in-house educational meetings.  Examples for solution (2) would be achievement of P4P objectives, use of screening tools which can be charged for, and making additional patient appointments to monitor chronic disease (this latter profitable only if the practice is not working at highest capacity already).  Another example would be convincing an insurer to pay higher rates in light of improved quality.  Examples for solution (3) would be using encounter templates and questionnaires that would improve visit efficiency at the same time as improving quality. 


Clinical quality measurement has a long history in research, but few examples applicable to a practice on a routine basis.  As scientifically trained clinicians, we know the importance of measurement to ensure validity of accomplishment; as practice managers, we know its importance for motivation of personnel.  If the practice should adopt laborious data collection, however, this allocation of resources could restrict QI efforts to only one or two objectives for years, and become so burdensome as to sink the whole QI enterprise.

If the practice could be nimble and inventive, however, measurement would not bog down progress.  The goal, after all, is improvement in a single practice, not scientific proof of replicability to other practices, nor achievement of statistical confidence levels.  Some QI steps can prove their worth on a prima facie basis simply by their existence – for example, instituting a regular channel of communication with a social service department, thus improving referrals.  Other improvements may be obvious – for instance, the dwindling height of the pile of pending laboratory reports, or the newly-empty waiting room as check-in procedures are improved.  Some will have intermediate measurability – for example, using P4P biomedical objectives as generated by either practice or insurer, but in the latter case rechecked by the practice.  Another example would be measuring total visit times (time from patient check-in to check-out) using appointment software along with other in-office adaptations. Subjective measures may have validity – for instance, increasing spontaneous compliments from patients, improved scores on patient satisfaction surveys, and the feelings of clinicians about their work.  Some objectives will be stubbornly immeasurable – for instance, the performance of clinicians with non-stereotypic complex diagnoses.  Importantly, the practice needs to resist the impulse to improve only the measurable: “Everything that counts can’t be counted, and not everything that can be counted counts.”


In this welter of considerations and absence of a standard approach, how can a practice proceed?  The important first step will be for the practice formally to accept the QI challenge, and to appoint a QI leader who is apportioned time and payment to do the job.  The leader then needs to start planning, involving practice members widely to ensure that the results aimed at are highly valued. The planning process will be constant and reiterative, as values, objectives, interventions, organizational feasibility, and financial reward are constantly balanced and rebalanced. The hope would be that not only the agenda, but also the process itself would initiate or strengthen a Culture of Quality in the practice, where quality is not in the back of the mind, but a constant preoccupation.  This culture will in itself breed success, perhaps in unforeseen ways.

The resulting QI agenda, whether narrow or wide, will need to generate objectives from quality definitions one, two, and three; choose SR and/or PE approaches for each objective; and find ways to overcome obstacles one, two, and three. Table One gives some idea of the variety of objectives a practice could consider, and the criteria by which they could be judged.

Practices are so different that it is clear that agendas will need to be very individualized. Emphasizing PE or SR would depend on a practice’s self-assessment – for instance, are strong clinicians hampered by disorganized processes, or are clinicians in need of stimulation, or both?  Likewise, is organizational capacity so low that only PE strategies are possible?  For all practices, it would make sense to include initially some “low-hanging fruit,” since initial successes are important, as well beginning on longer-term objectives.  P4P objectives would often be the “low-hanging fruit.”

A Concrete Example

To help in visualizing possible practice agendas, here is an example of a two-year initial agenda with some explanatory detail. 

Several P4P objectives successively, according to offers of reimbursement by payors, probably starting with immunizations.  Each objective would have to be accepted according to the extent of necessary effort – in other words, will it be worth the work.  PE would include educating clinicians on “immunization opportunity visits.”  SR would be clear charting on a separate chart page, medical assistant reviews, and patient recall from billing data. The financial obstacle would be mitigated by health plan payments, and perhaps additional visits from the recall program. Conceptualizing and organizing the effort would be the biggest obstacle, but once done, the same organizational pattern could be used on further P4P projects. Measurement would be by the health plans – as checked by chart review by the practice, as necessary.  These objectives could be “low-hanging fruit.”
Production and implementation of templates.  These practice aides (SR interventions) promise to reduce variability and, as a joint product, to improve efficiency. They change acquiring and imparting standard information from oral-aural to the more efficient and effective utilization of questionnaires and handouts. Pediatric examples would be a questionnaire for a first visit for wheezing, well visit encounter forms for various age groups, and handouts for anticipatory guidance.  Adult examples would be use of routine questionnaires for alcohol use, tobacco use, and depression; and introduction of diabetic flow sheets.  

They would require some organizational leadership and cooperation to produce among the clinicians, and while they would take time and effort of no small degree, there would be little out of pocket costs.  If well designed and accepted by clinicians, implementation would not be difficult, since it would consist merely of inserting printed forms into the visit.  The net financial result might be positive, since templates such as these make a clinician’s work easier and more efficient.  Measurement requirements to assess the innovation would be minimal, since there would be prima facie decrease in variation.  But unquestionably these projects would require significant thinking, cooperation, time, and effort.  Predicting ultimate profitability would not be certain.

Total visit time (TVT) reduction and improved telephone performance.  These efforts would aim at the second quality definition, service.  Since this would involve primarily non-clinical staff, they could run parallel to clinical efforts without overstretching the practice.  Interventions would start with measurement of TVT, and telephone surveys of service conducted by fake patients.  SR interventions would be procedure outlines for staff, supplemented by training (PE).   These would require extensive leadership and design, and possible information technology consultation.  Improved service will be obvious to patients, and thus more patient volume might result.

Periodic case conferences.  In-house conferences where clinicians trade their puzzles, triumphs, and even failures, promote learning and quality consciousness.  Another version calls for mutual chart review, where the clinicians read each others selected charts, and discuss not only the cases but also the documentation and workups.  Conferences on “difficult patients” with experienced guests or group members can improve empathy (definition three).  A PE effort, it requires little organizational effort, entails little financial input and no risk.  No objective measurement is possible.

Encouragement of participation of one practice member on hospital QI committee.  Another clinical, PE action, this would educate and perhaps inspire one clinician at little cost, although one must recognize that 12-24 hours or more spent per year is not insignificant.

Establishment of personal liaison to one social service worker.  This would be an easily accomplishable step requiring initial telephone inquiries and perhaps a lunch meeting.

These steps would constitute an ambitious program.  Clearly, some projects would end quickly and could be replaced by others, while others would take several years of effort.  If this much activity were embarked upon, one would expect that a major change in the zeitgeist of the practice would result.

The Electronic Health Record

A note is needed about EHR’s, since EHR proponents frequently assume that much of QI would be accomplished it information technology such as EHR.  It is true that many QI steps discussed here as paper-based technologies might be more elegantly accomplished with EHR’s.  EHR might also provide easier chart review, easier use of patient registries, easier reminders of clinical practice guidelines, easier communication among providers, etc. EHR would also force reluctant clinicians to use tools that they might ignore were they only on paper.   No one doubts that EHR is in the future.

Introducing EHR to an office, however, is expensive and risky, and the financial upside very limited.   The contribution of EHR to quality remains in question.  Many desired computerized functions (patient registries, immunization lists, etc.) can be provided by a computerized billing system, already standard equipment in most offices.  Critics wonder why “the doctor should be asked to invest in medical record systems when primary systemic beneficiaries are elsewhere?” It is doubtful that one marginal business case (EHR) will facilitate another marginal business case (QI).  Since most QI steps will be translatable to EHR when they appear in a practice, and since much of what EHR do can be done with paper charts and billing systems if not so elegantly, indeed since QI beforehand may actually facilitate QI after EHR implementation, QI should not wait on EHR’s, and above all EHR’s should not be viewed as a substitute for QI.


While the need for QI is clear, the obstacles are formidable, and the incentives beyond idealism appear weak.  There is no specific road to follow, and few practical measures. Many objectives, however, can be reached rather easily.  It seems clear that if quality is to improve, practices must consider issues of their own organization very seriously.  Despite the current climate of cynicism, medicine has historically been a repository of idealism.  It does appear that if a practice is idealistic enough to want to pursue QI, and if it adopts an intelligent plan, it need not suffer for having chosen to do so, and indeed, might benefit as much as its patients for the effort expended.

1 comment:

  1. Still very relevant, Budd! Your comments on P4P were prescient