Process Driven Methods in Diagnosis and Treatment




(1)
Department of Neurology, Wake Forest University School of Medicine, Winston-Salem, NC, USA

 



Abstract

This chapter explores processes for establishing diagnosis and guiding treatment in a dependable manner. It introduces the idea of creating a workflow to visualize, monitor, measure, and improve the steps involved in diagnosis and treatment. It presents variance reduction methods like six sigma, checklists, and describes potential applications to medicine and neurology. A process driven method can make matters more efficient by providing opportunities for integrating operations and running many processes in parallel. The benefits of integrated, multidisciplinary clinics like the Amyotrophic Lateral Sclerosis (ALS) clinic are presented. This chapter also borrows ideas from the field of product lifecycle management (PLM) to develop similar principles for managing diseases across their lifecycle called disease lifecycle management (DLM). In a manner similar to prior chapters, the principle is presented first followed by medical case examples drawn from daily practice. This chapter forms a bridge that extends the journey from the first part of the book that emphasizes decision making and diagnostics into the second half that explores improving treatment.


Keywords
Systems engineeringValidationVerificationProcess driven methodsOpen problemsHypokalemic weaknessAcid maltase deficiencyContinuous engineeringMyasthenia gravisPlatform-based product developmentMultidisciplinary integrated clinicsAmyotrophic lateral sclerosis (ALS) clinicProduct lifecycle management (PLM)Disease lifecycle management (DLM)Chronic inflammatory demyelinating polyneuropathy (CIDP)Statistical controlVariationsControl chartsSix sigmaOrder setsDiagnosis-related processesChecklistsImmunosuppression checklistsIVIG checklistsTroubleshooting checklists



Introduction


This chapter looks at the application of manufacturing principles to healthcare. Manufacturing enables consistent duplication and delivery of the same product to different customers without variation. A sequence of predefined steps is necessary in creating the final product which constitutes the assembly line. The ideas and science behind it form the basis of management and engineering—how do we organize production efficiently to guarantee the highest quality, least defects and variations and minimum cost? The Oxford dictionary defines a process as “a series of actions or steps taken in order to achieve a particular end” [1]. This chapter will borrow established ideas from management and engineering and apply it to healthcare delivery by creating a process. For the most part, these initiatives are zero cost and serve to adapt successful manufacturing industry principles for healthcare processes; therefore no investments in equipment or drugs are necessary.


The Benefits of Defining a Process


The world of manufacturing is driven by tangible inputs and outputs—such as a company manufactures a certain number of cars of a certain quality or produces a certain number of metric tons of steel. The input is tangible and well defined. For the example of a car plant it is steel, rubber, plastic, paint, modular components (such as engine assemblies, gearboxes, etc.) and the output is finished cars. Once the process is defined, metrics can be applied for input and output. For a particular model, annual production takes x millions tons of steel, y gallons of paint, etc. for manufacturing z number of cars. For each input and output, a sequence of steps or a “manufacturing process” is defined to convert raw materials into finished cars. A process itself leads to subprocesses such as manufacturing a gearbox which feeds into the larger automobile manufacturing process of including the gearbox as a component of the finished automobile. Between the raw input and final finished output, there are a number of intermediate stages where products pass from one step of the manufacturing process to the next, often in an assembly line to yield the final product. The following are several advantages to formulating a process for a product:

1.

The input is well defined.

 

2.

The output is well defined.

 

3.

Intermediate stages are well defined.

 

4.

Metrics for the process can be defined.

(a)

Processing times for each step of the process.

 

(b)

Labor and technology requirements.

 

(c)

Quality can be defined.

 

(d)

Costs can be defined.

 

 

Once process metrics are defined, improvements can be planned. Such improvements can be radical, including rethinking the product or process itself or incremental. At some level, striving for improvement constitutes innovation. An example of radical rethinking for the above toy example would be to do away with the current machining process for manufacturing gears and use 3-D printing instead for manufacturing gearboxes. Incremental innovation usually involves refining the process or making minor modifications to finished products for lowering costs or improving quality. For example, given the same levels of manpower and shop floor equipment can we reorganize how we do things to increase output of gearboxes by 15 %? This innovation will be discussed further in Chaps. 8 and 9.

A process therefore is like an algorithm in computer science. It is a sequence of steps that must be followed from beginning to end for each iteration. Every car, every ounce of steel has been through the same steps in manufacturing. While it is difficult to say there is only one way to do things, once a process is defined and becomes measurable, it can lead to continuous experimentation for better ways to do things in terms of the defined metrics. As Lord Kelvin, the famous nineteenth century physicist once said “if you cannot measure it, you cannot improve it.” Therefore, innovation applied to processes leads to gradual maturation and a path towards the best way to do things.


Systems Engineering and the Traditional V Model


Systems engineering is described as “a methodical, disciplined approach for the design, realization, technical management, operations, and retirement of a system. A system is a construct or collection of different elements that together produce results not obtainable by the elements alone. The elements or parts can include people, hardware, software, facilities, policies, and documents [2].” Systems engineering is interdisciplinary by its nature [2]. It is concerned with a system’s performance, interactions of its subsystems with each other, interactions with the user, and other interacting systems [3].

The classic systems engineering model is called the V model as shown in Fig. 7.1.

A327463_1_En_7_Fig1_HTML.gif


Fig. 7.1
The traditional V model for systems development. Adapted from [2, 3]

The traditional model assumes a sequential development process and starts at the top of the left arm of the V. It proceeds stepwise from left to right along the arms of the V. The first step is defining and understanding customer requirements. Based on customer demands and understanding of needs, a system specification is arrived at—termed validation. Validation refers to confirming that the system being developed conforms to the user’s needs and expectations [3]. The system specification is implemented in a top-down manner by first specifying the high level design of the system. For the example of an automobile, high level design would include decomposing system development into platform, engine, transmission, braking, and electronic control systems. Subsequent development proceeds by systematically decomposing each high level system into component subsystems (detailed design) and ultimately to component level detail [2, 3]. The idea is to attain system decomposition into component subsystems which can be implemented in parallel to speed product development [3]. Each step of the process requires detailed documentation since it is expected that the personnel may leave the project along the development lifecycle.

The right branch of the V is where components and devices are initially verified before being integrated into subsystems. Verification refers to determining that the system being designed meets all the predefined specifications. This step confirms the design is robust, maintainable, and each component behaves as expected [3]. Individual components/devices are then integrated into subsystems. Each subsystem subsequently undergoes rigorous verification before integration into the final system. Verification is then performed on the completed system to ensure that the entire product conforms to requirements specified in detailed design stage of development [3]. This is followed by validation to ensure it meets the customer’s requirements.

Understanding customer needs and getting the right system specification is extremely important prior to proceeding with system development. This is because making changes after detailed design is complete is extremely difficult since it involves redoing a lot of the development work to meet the new requirements. Therefore, the traditional model lays great emphasis on the first step of understanding customer needs and specifications prior to system design with a sequential implementation as a consequence.


The Lack of Processes in Healthcare


A common theme in many healthcare problems is the wide variability in patient experiences. The elapsed time between onset of symptoms and final diagnosis and instituting treatment varies from days to weeks to several years. Events such as a clinic visit or a laboratory test happen haphazardly, there is a period of some activity followed by quiescence. Similarly, costs are also extremely variable for the same diagnosis and comparable outcomes. This chapter will explore methods for creating processes which will have the advantages listed in Section “The Benefits of Defining a Process”.


Defining Patient Processes and Creating a Virtual Assembly Line


A similar V-shaped model can be created for healthcare delivery. To a great extent, matters happen in isolation in healthcare with all too great an emphasis on the physician. The first step is to define a healthcare process and create a virtual assembly line or a V model for each patient. No two people have the same disease or require the same treatment making it difficult to create standardized flowcharts in healthcare. However, a process can be created for each patient to improve healthcare delivery. The following steps can be identified:



  • Step 1: Define patient symptoms and needs: This step is analogous to “concept of operations” in systems engineering. This step involves two major objectives:

    (a)

    Understanding patient symptoms to direct diagnosis.

     

    (b)

    Preliminary understanding of patient requirements and needs to define initial objectives of care.

     


  • Step 2: The next step involves subsequent information gathering performed during the clinic visit—clinical examination, review of medical records as the necessary background for clinical decision making.


  • Step 3: Clinical Diagnosis step: This step involves formulating a working diagnosis. This step can take any of the forms discussed in prior chapters such as Fault Tree Analysis (FTA), Graphical Methods, and Probabilistic methods. Additionally, the problem maybe cast in a Byzantine framework for decision making in the face of byzantine faults as discussed in Chap. 6.


  • Step 4: Direct appropriate diagnostic testing: In this step, the clinician directs diagnostic testing based on possibilities identified in step 3. If the diagnosis is validated he proceeds to the next step, if rejected he proceeds with alternate hypothesis (such as other hypothesis identified in the FTA or graphical method).

As shown in Fig. 7.2, this step is iterative and must be traversed repeatedly until the correct diagnosis is established. The speed with which this loop is iterated and diagnostic possibilities examined varies depending on the severity of the problem, as will be discussed later.

A327463_1_En_7_Fig2_HTML.gif


Fig. 7.2
Creating a patient process analogous to the systems engineering V model





  • Step 5: Diagnosis confirmed: In this step, the clinician establishes both the diagnosis and the degree of confidence in the conclusion. This helps direct the choice of treatment. As described in Chap. 6, it is useful to define the working diagnosis in the byzantine framework. Treatment can be selected based on the confidence-cost/risk matrix in Fig. 6.​4 of Chap. 6.


  • Step 6: High-Level Treatment planning: In this step, the clinician and patient can discuss an appropriate treatment plan that conforms to the extent possible with patient expectations identified in step 1. Patient goals, intermediate milestones are identified and established in this step. Examples include “be able to walk with a walker in 2 months and with a cane in 4 months.” In this step, the decision making is along the lines of broad generalities such as “Immunosuppression” and “Physical Therapy” or “Symptomatic Treatment” instead of immunosuppression for “myasthenia gravis”.


  • Step 7: Detailed treatment planning: This step involves defining the low level fine details involved in formulating treatment. It involves performing a preliminary system safety assessment (PSSA), followed by system safety assessment (SSA). Functional hazard assessment (FHA), failure modes and effects analysis (FMEA), and particular risks analysis (PRA) are performed to guide therapy. Methods to perform these are described in detail in Chaps. 1, 2 and 4.


  • Step 8: Verification: In this step, treatment delivery is verified, are medications being taken as directed? Is therapy focusing on the right goals and objectives? At this stage side effects, anticipated and unanticipated are reviewed and mitigated. This step also includes appropriate therapeutic monitoring for dose corrections such as INR monitoring for warfarin.


  • Step 9: Monitor response to treatment: This is a strong validation step which affirms the veracity of the first 8 steps and enables corrective action if anticipated goals are not met. If there are shortfalls these imply:

    (a)

    Imperfections in diagnosis: this is addressed by going back to step 4 and performing FTA, graphical methods or byzantine generals framework analysis to improve the diagnosis.

     

    (b)

    Imperfections in treatment: Treatment can be refined by using the Plan-Do-Check-Act (PDCA) cycle which will be discussed in Chap. 8.

     


  • Step 10: Collect and analyze data for Quality Assurance and Knowledge building: This step connects back to step 1 where the process can be iterated again for a similar clinical condition with the benefit of lessons learnt from prior experience.

The systems engineering approach yields a few important lessons:

(a)

Successful care delivery is a multistep process which must work harmoniously. It is greater than the sum of its parts.

 

(b)

The process is dependent on different skills and expertise at different steps of the journey. Therefore, it must necessarily be more than one person, the physician and must involve teamwork.

 

(c)

While the steps involved are very similar between different clinical problems, the speed with which the cycle must be traversed differs greatly depending on the acuity of the problem. Severe, life-threatening problems would require rapid process times (in minutes to hours) while for chronic outpatient problems, the process can be traversed over days.

 

(d)

The model allows for factoring time, cost, and expertise at each stage of the process. During early stages, the emphasis is on the left arm of the V since most of the effort needs to be expended in determining the diagnosis. In later stages, the emphasis shifts to the right arm of the V since the diagnosis is firmly established and it is refining treatment that is typically important. This model provides a foundation for transitioning to the Disease Lifecycle Management (DLM) model which borrows ideas from Product Lifecycle Management (PLM).

 

The traditional V model finds great applications in single provider settings, where diagnosis and treatment is essentially dependent on one physician or a group of physicians. In the case of clinical problems involving more than one type of physician (such as a multisystem disease with cardiac, renal, hematological involvement) or where expectations are dynamic and goals need continuous refinement, newer systems engineering models are advantageous and they will be described later with case examples. Detailed implementation of steps in the V will be studied further in the Reliance Microplanning method in Chap. 9.


Defining Processes Based on Time


Based on time, complexity, and number of systems involved, problems can be labeled as follows:

Open problems: These are problems where a diagnosis has not been established. Based on acuity, severity, and duration these are classified as “Open-Chronic” and “Open-Urgent/Acute”. Most of the cases discussed in prior chapters are examples of “Open-Chronic” where the physician has time (measured in days to weeks) to guide thoughtful analysis and testing. In these cases, the V can be traversed in weeks or months. “Open-Acute/Urgent” problems are the most dangerous, since the patient’s condition is rapidly changing and unless appropriate action is instituted in a timely manner, there is potential for irreversible injury and loss of life. Such problems may need correction of life-threatening anomalies to be instituted before a root cause can be established and treated. Examples include rapidly declining ventilatory function which requires immediate intubation prior to determining if this is from myasthenia gravis or Guillain-Barré Syndrome (GBS).

Open–multidisciplinary: These cases involve more than one specialty, therefore a multidisciplinary approach is necessary. In other cases, the diagnostic workup threw open a Pandora’s Box of frequently overlooked information which requires partnership and assignment of responsibility for appropriate treatment. This classification is important for improving dependability in healthcare. “Open–Multidisciplinary” are cases where the devil is in the details and a small little thing buried in a test report which means very little to the physician who ordered it, probably will snowball into the most important thing in the years to come.

Closed: All aspects of the presenting symptom(s) have been addressed and no further diagnostic workup or treatment is necessary. The aim of traversing the V is to go from open to closed status.


Case Example 1: Open-Acute—Urgent Problem


JH is a 35-year-old male, s/p kidney transplant who presented to the hospital with diarrhea and dehydration. Medical comorbidities include hypertension and diabetes mellitus. During the course of his admission, he was treated with supplementation of IV fluids, electrolytes, and monitored closely for renal function. He awoke one morning with weakness of the left arm. He also had mild left shoulder pain and did not feel there was any difference in sensation. On examination, he had diffuse weakness involving the left arm involving all muscle groups with normal reflexes everywhere. Sensory examination was normal.

The history and physical examination did not yield an immediate solution. The weakness was too diffuse to be attributed to the commonest reason—radial palsy. Given the sudden onset, diffuse weakness and risk factors of kidney disease, high blood pressure, and diabetes mellitus the most pressing concern was for a stroke. An urgent MRI Brain was requested. The patient was beyond all windows for IV or intra-arterial tissue plasminogen activator (tPA). The patient was initially seen in the afternoon around 3:00 p.m. The MRI Brain was planned within the next 2–3 h as soon as scheduling permitted. If this was negative, the plan was to proceed with an MRI of the Cervical Spine to see if this was due to lesions of the spinal cord like transverse myelitis. If this was negative as well, the plan was to perform an EMG to look for brachial plexus lesions. The problem was classified as Open-Acute/Urgent which necessitated a solution had to be found in hours.

By 7:30 p.m., MRI Brain was performed. The MRI Brain was reviewed later that evening and was found to be normal. No evidence of stroke was found. Proceeding down the algorithm mentioned above, MRI C Spine was ordered urgently. While waiting for this, a thorough review was performed, including consideration of the rare and obscure (West Nile infection, CMV infection etc.) which though possible given the immunosuppression were obviously not the cause here. MRI C Spine was completed around 10:00 p.m. and was normal as well.

Since no solution had been apparent by 10:00 p.m., a second examination was performed of the patient around midnight, approximately 8 h since initial examination. By now, the patient had woken from a nap and was paralyzed below the neck. Breathing, eye movements were normal. Curiously speech and sensation were normal and he had no urinary or fecal incontinence. Since no firm hypothesis was formed, reevaluating the entire data looking for new clues became the priority. An intense review of all data to date performed around 1:00 a.m. showed a fact that had been overlooked in the afternoon, a low potassium level of 2.8 meq/L. An urgent repeat level was ordered which showed it had slid lower to 2.0 meq/L with the potential for life-threatening cardiac arrhythmias. This was not suspected or sought aggressively initially since low potassium is not a common consideration in focal weakness.

The solution had been determined but only just in the nick of time. An urgent EKG showed findings of severe hypokalemia such as U waves and frequent ectopics. The patient was immediately transferred to the ICU. Potassium was urgently replenished with complete recovery. The root cause of the problem was identified as loss of potassium from diarrhea and changing acid base balance.

This case would have been fatal if the hunt for the solution had been delayed until next morning. Therefore, a problem status has to be defined and the V model has to be intensely applied until the problem is solved within the acceptable timeframe. As discussed above, approaching this as a process has the advantage of imposing intermediate processing times for each step of the diagnostic approach with alternate ideas being explored in a time bound way driven by the urgency and importance of solving the problem.


Case Example 2: Open-Chronic


WH is a 60-year-old male with symptoms of bilateral arm weakness. He did not remember when the problems started, it could have been many years ago, but it progressed very slowly. The first symptoms he noticed were difficulty retrieving objects from shelves. He had no difficulty with closing his eyes, speaking, chewing, swallowing, breathing, or walking. He had no problems with sensation. On physical examination, he had severe weakness about the shoulder. The mildest movements would lead to displacement of the shoulder blade so that the deltoid muscle was rendered mechanically ineffective. The deltoid muscles themselves showed only mild weakness. The serratus anterior and rhomboids were severely wasted and weak with prominent scapular winging. The face, forearm, hands, and legs were normal. Sensation was normal. No diagnosis had been arrived at to-date. FTA performed for this case is shown in Fig. 7.3. Based on this analysis, the prominent scapular winging an initial working diagnosis of Facioscapulohumeral dystrophy (FSHD) was considered [4]. Neuropathic etiologies which can cause scapular winging, (usually in a scapuloperoneal pattern) include Davidenkow’s syndrome [5]. A gene test for FSHD (D4Z4 repeats on chromosome 4q35) was requested [6], patient was provided a return appointment in the next 4 months as results of diagnostic testing were obtained. The patient was also set up for EMG.

A327463_1_En_7_Fig3_HTML.gif


Fig. 7.3
FTA for Case Example 2. *IBMPFD: Inclusion Body Myositis with Paget’s disease and frontotemporal dementia

The patient underwent an EMG performed by another physician which confirmed a myopathic process. Genetic test for FSHD was negative which can happen in ~2 % of cases [4]. This necessitated going back to steps 3, 4 to examine alternative hypotheses from the FTA and direct further testing. Since the gene test was negative, a muscle biopsy was requested based on raw EMG data. Though clinically weak, commonly biopsied muscles such as biceps were reported to be normal on EMG. Based on EMG description, the findings in the infraspinatus were interpreted to be reasonable for biopsy. The biopsy was however indeterminate since the sample was essentially end stage muscle. Since the patient lived far away, the muscle biopsy was performed concomitantly with blood testing for Acid Maltase (Pompe’s disease) activity level, which is considered a rare cause of scapular winging [6]. Pompe’s disease is classically described to produce limb girdle weakness and respiratory weakness which was not the case here [6]. The Acid Maltase activity level was low confirming this to be an unusual presentation of Pompe’s disease. Pompe’s disease can cause respiratory weakness but is treatable with enzyme replacement therapy, therefore it was a surprising outcome [7]. While Acid Maltase deficiency itself is a very rare disease even for most skilled neuromuscular clinicians, this presentation of it is even rarer. Application of this rigorous process yielded a solution instead of attributing his problem to a variation of FSHD. Pulmonary Function tests showed moderate diaphragmatic weakness.


Solutions Process

The diagnosis was confirmed by genetic testing to be Pompe’s disease. At this stage, the transition began to the right arm of the V.

The next steps of High level treatment planning, treatment goals, including supportive care were discussed in detail with the patient. This is vital since Pompe’s disease causes severe ventilatory weakness which can shorten life expectancy and is treatable. The patient transferred care to another clinic for the same.

We will jump ahead to the Quality Assurance and building knowledge part. The following errors were observed in this case:

1.

Poor choice of muscle biopsy due to poor muscle selection. The test yielded no useful information and left the patient with a bill which he is still struggling to pay off. The root cause of this was poor interpretation of EMG data since this was done by another clinician. This could have been avoided if a discussion and review had been performed with the physician performing the EMG so that he could advice on the best choice for biopsy. Therefore there was failure of communication and teamwork.

 

2.

Incomplete EMG: The thoracic paraspinal muscles were not studied in this EMG. The thoracic paraspinal muscles can show a peculiar pattern of electrical abnormality called electrical myotonia which could have tipped the clinicians in that direction [7].

 


Lessons from WH, Plans for Solution Improvement



1.

The best muscle to biopsy should always be stated explicitly in the report. If this is not the case, always discuss with the physician performing the EMG which is the best muscle to biopsy.

 

2.

In all cases where primary muscle disease is suspected, perform EMG of the thoracic paraspinals. This is to look for electrical myotonia and also to look for very proximal muscle involvement which can be seen in inflammatory myositis like dermatomyositis and polymyositis.

 

The lessons learnt from WH helped improve the diagnostic process for patients with similar problems because this experience was incorporated into future cases and no future negative biopsies were encountered.


Open-Chronic Ancillary Multi-Specialty Problems and Related Processes


This section discusses creating ancillary processes for unexpected information encountered during the diagnostic process in steps 3 and 4 which are not the core area of the physician who initiated the process. Teamwork in the care of complex multisystem patients is discussed separately in subsequent chapters, this example guides a process driven approach to information unexpectedly encountered while traversing the V.

Advances in imaging technology and laboratory diagnostic medicine have created immense amounts of ancillary information which are discovered serendipitously during the course of searching for something else. A major problem has been the lack of a rigorous disciplined approach to handling such unexpected information when they are not the focus of the ordering clinician. Numerous examples abound in virtually all areas of medicine with radiology being the common denominator in most instances. Successful modern management in industry involves defining problems and subproblems and assigning responsibility appropriately. Such details must not be omitted and must be addressed by applying the correct skills. For each thread of unsolved information, a process must be defined and responsibility fixed so that a problem does not get neglected and become incurable, especially in those happy instances when a long lead time would have benefitted treatment. Assigning responsibility is key. The following example illustrates this principle:


Case Example 3


MA is a 65-year-old male smoker with symptoms of double vision, fatigue, and weakness for several months. Strokes, aneurysms were excluded by normal MRI Brain scans. He was diagnosed with myasthenia gravis. The diagnosis was fairly straight-forward, blood work was positive for the antiacetylcholine receptor-binding antibodies (AchR-Binding antibody) which established the diagnosis. Since a percentage of cases are related to a tumor of the thymus gland located in the chest, a CT scan of the chest usually follows this diagnosis. In MA’s case, it was anticipated that his treatment would involve suppression of the immune system with steroids such as prednisone. For this reason, precautions were taken to check vitamin D levels, thyroid function, establish whether he was diabetic or not and perform a skin test for tuberculosis along the lines of FMEA discussed in Chap. 4.

The CT chest was negative for thymoma, but showed pulmonary nodules which are a frequent finding in CT scans of the chest as scanners became better. Blood testing showed he was already diabetic without knowing it. Thyroid function testing revealed hypothyroidism. Additionally, vitamin D levels were low. Given this constellation of abnormalities, responsibility and follow through were defined for the subproblems:

(a)

A copy of the CT chest report was provided to the patient identifying the size and location of the nodules. A repeat scan was recommended by the radiologist in 6 months to see if there was any growth suspicious for lung cancer given his smoking history. This responsibility was transferred to the primary care physician and patient so that it was not overlooked in the course of treatment of his myasthenia gravis.

 

(b)

The constellation of diabetes mellitus and hypothyroidism was treated by referral to an endocrinologist. The anticipation that he will need steroids which will worsen diabetes considerably and that be factored into treatment planning was made and responsibility assigned. By the time of his return visit in 2 months, both these problems had been well treated. This enabled safe use of prednisone for this patient with minimal complications of worsening diabetes since the sugars were well controlled and the treating endocrinologist had formulated contingencies for perturbations in sugar control with the use of Prednisone at a moderate dose for 2 months. Addressing thyroid abnormalities helped with overall metabolism, decrease in fatigue and improved exercise ability which made myasthenia gravis easier to treat.

 

Therefore, what started out as one symptom had multiplied into four problems—lung nodules, diabetes mellitus, hypothyroidism, and myasthenia gravis. Each required meticulous attention and follow through which was beyond the scope of one person. Defining a problem list where each of these entities is enumerated with an appropriate physician taking responsibility for each problem was defined. In a sense, it is similar to the graph theoretic structure described earlier except the idea is not to define an edge between the nodes but to establish a thread of responsibility and follow through. The advantage with defining such problem lists is that every subproblem gets labeled and gets adequate attention, expertise, and follow through. The recent requirement for creating, maintaining, and updating such problem lists is a step in this direction.


Continuous Engineering and the Integrated Workflow Model


The traditional V systems engineering model suffers from one drawback. It is very sequential. It lays extreme emphasis on understanding customer needs and defining requirements prior to initiating detailed design. The model performs poorly when customer needs are evolving—this can happen as a consequence of customers not knowing what they want or being unclear about their exact needs and from changing market conditions. In recent times, the traditional V model has been supplanted by more integrated models which see customer needs as flexible and therefore need refinement throughout the system design process. This model is referred to by many different names in different industries, one term that is applied is continuous engineering [3].

Continuous engineering involves “continuous verification and validation throughout the product development process [3]”. The basic steps are similar with the left side involving definition and decomposition and the right involving integration and validation. The different steps of the V happen massively in parallel and iteratively as system design is continuously refined and changes implemented [3]. Given constantly evolving needs, system design changes, continuous verification and validation, continuous engineering needs very high transparency, teamwork, knowledge collaboration, and excellent communication [3]. Emphasis is placed on modular designs which are interchangeable across product lines. Reusing components, software, knowledge, and assets during different product development cycles helps manage cost and time [3].

An integrated approach has been well adopted in many industries, especially the auto industry. In earlier times, model changes happened sequentially—first the design, then the engineering, then investments in tooling, production planning and finally sales. This led to an all too common problem—conflicts between stages in development and expensive redesign. A design may not be the optimum one to manufacture for time or skill reasons which would then require redesign and slowing of product development. Automakers have increasingly relied on integrated approaches where downstream processes such as manufacturing, sales are built into the design stage itself and a car is designed to manufacture and sell. This can be carried to a further level where automakers have increasingly adopted component sharing across product lines which has reduced product development costs considerably and led to benefits from economies of scale. Additionally, companies have moved from building a single product with firm specifications to a family of products spanning a spectrum of specifications catering to a diverse need. This has considerably reduced lead times or cycle times in the auto industry where model replacements now happen on average every 4 years from the usual 8–9 years in the sequential approach. This is referred to as platform-based product development. Formally, a platform-based approach is described as “a collection of the common elements, especially the underlying core technology, implemented across a range of products” [8, 9].

The integrated, platform approach allows for ideas from all levels of the process to be applied at the very beginning and may potentially resolve conflicts before they happen. In this chapter, we borrow the term platform approach since it brings together multiple different skills in a platform to deliver the maximum in one clinic visit similar to what happens in industry. The interested reader will find a wealth of information on integrated approaches to healthcare in [10].

Such an integrated approach can be applied to healthcare to potentially speed diagnosis and treatment. An example is the general neurology clinic which is usually run in the sequential workflow model. In the classic healthcare model (especially for large teaching hospitals), a patient sees a consultant who then orders some tests, imaging studies, seeks the assistance of another consultant, and finally based on putting all these together comes up with a treatment. Medical records are sought and reviewed when they are received. The treatment itself may take the form of physical therapy or occupational therapy with further stages in coming to fruition. Adopting the platform approach led to the following solution method in a general purpose clinic.





  • Step 1: Review the problem—the specific question asked by the referring physician which may take the form of “Evaluate for nerve damage”, or “Evaluate for muscle damage,” etc. If the question is not well formulated, make an a-priori hypothesis. This is akin to understanding customer needs before even seeing or talking to the customer.


  • Step 2: Assess acuity and offer an appointment. If EMG testing is necessary, offer an appointment on a day when the two—a clinic visit and an EMG visit can be integrated.


  • Step 3: Seek pre-certification from insurance companies so that patient is approved for clinic visit and EMG and the two can happen in a smooth sequence.


  • Step 4: Plan the range of blood tests that maybe necessary—special blood work for immune system disorders or genetic disorders.


  • Step 5: Plan for additional consultant expertise—does the patient need a biopsy of a muscle or nerve? Select the one which is anticipated and reduce lead times further by requesting the required appointment or procedure well before seeing the patient. This may not be needed after the clinic visit and EMG based on the results obtained, therefore such an appointment can be flexibly overbooked in most instances.


  • Step 6: Split the visit into two parts, the first half dealing with diagnostic history, physical examination proceeding straight onto the EMG part of the study. Integrating the two in the same room saves a lot of time and helps make the process parallel. Once this is done, the second half can be reserved for discussing potential diagnosis and treatment plans.


  • Step 7: Perform blood work; patient sees the surgeon for potential biopsy or other consultant whose services are requested.

This method is appreciated by patients who sometimes wait several months for an appointment and drive several hundred miles. The classic method disconnects all these processes, sees each step in isolation and requires many visits to the hospital/clinic separated by weeks and at considerable cost and delay. The platform approach therefore involves problem anticipation, solution planning in a manner similar to the microplanning method which will be described later.


Integrated Clinic: Amyotrophic Lateral Sclerosis Clinic


One of the best examples of the integrated platform approach is the ALS (Amyotrophic lateral sclerosis) clinic. While there is no specific treatment for this condition, once the diagnosis is confirmed, patients are referred to the ALS clinic which is dedicated to the long-term care of such patients. Riluzole has a modest effect at slowing down the progression of the disease. Since the disease is frequently relentlessly progressive with involvement of breathing and swallowing muscles at different stages of the disease, great teamwork is needed to ameliorate deterioration and maximize quality of life. The classic sequential approach in a debilitated patient is very difficult given the physical and logistical difficulties in going from one appointment to the other, sometimes over great distances. The care of the ALS patient is discussed in [11, 12]. The ALS clinic uses an integrated or platform approach to deliver very high quality, dependable, supportive care [11]. This model can be adopted in other clinics to deliver better care and dependability. The clinic seeks to address the greatest challenges faced by patients during each visit. Figure 7.4 shows the integrated patient care model of the ALS clinic.

A327463_1_En_7_Fig4_HTML.gif


Fig. 7.4
Multidisciplinary ALS clinic, an example of an integrated systems engineering model. The ALS clinic can be seen as an integration of multiple subsystems such as speech therapy, physical therapy, nutrition requirements, respiratory assessment, and miscellaneous (blood work monitoring, eligible clinical trials, social aspects of care). An integrated approach permits these processes in parallel and integrates the effect of these subsystems

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Sep 24, 2016 | Posted by in NEUROLOGY | Comments Off on Process Driven Methods in Diagnosis and Treatment

Full access? Get Clinical Tree

Get Clinical Tree app for offline access