PROCESS IMPROVEMENTS

Translate Know-What Into Know-How

by Ed Chaplin, M.D., and Yoji Akao

Organizational change is a much more difficult challenge than we've been led to believe. In fact, a study of business magazines during the last few years led to an assumption that less than one-third of change efforts had an impact on improving competitiveness.

Some factors contributing to this relatively poor track record include human and organizational inertia, natural human tendencies such as hindsight bias and fundamental attribution errors, and breakdowns in translating targets into action. This article focuses on the latter--the challenges of translation.

Voice of the customer (VOC) methods and the tools of quality function deployment (QFD) suggest it is possible to capture 90 to 95% of customers' wants and needs by interviewing a dozen or so customers.^{1, 2}

Studies in the theory of inventive problem solving (often referred to by its Russian abbreviation, TRIZ) indicate 95% of the knowledge to solve most technical problems usually already exists in other industries.³

Why, then, are we not more successful in our efforts to improve organizational performance and create customer delight?

Do you remember as a child sitting in a circle for a game, sometimes called "telephone"? One child whispered a sentence into the ear of the next child, that child to the next and so on. What was the result after a complete turn around the circle? Often it was a completely different sentence.

When we use this example of human-to-human communication in our presentations to groups, many in the group reflexively laugh as they recall such experiences. Yet this same phenomenon happens in organizations every day.

Two Types of Knowledge

Translating what customers tell you or incorporating the technical know-how from another industry into routine practices is easily conceptualized but often difficult to actualize. Why? One reason is we have two distinct knowledge systems operating within our brains, and when we fail to distinguish them, we create fixes and deploy strategies that fail.⁴

Most of you have probably used a computer keyboard. Where is the "u"? Like most people, if someone asks you where the "u" is when you are not at the computer, you'll have to stop for a moment and visualize the keyboard. Yet, if you are at the computer, you instantly move toward the "u" whenever you need to use it.

Someone may ask you to spell a word. At first you cannot, but then you take a pen and write it out. Or you try to recall a phone number for someone but cannot, yet you go to the phone and dial it.

In each example the first attempt at knowing is conceptual, while the latter is tacit or procedural. Kujiro Nonkaka and Hirotaka Takeucho⁵ have defined these kinds of knowledge as "know-what" (explicit knowledge) and "know-how" (tacit knowledge). These two types of knowledge rely on two distinct memory systems within the brain.⁶ We encode and access each type of knowledge very differently.^{7, 8}

For example, many managers act as though explaining something conceptually readily translates into knowing how to do it. But the majority of our everyday activities are initiated automatically before thought.⁹ We grasp concepts via our communication with others, but encode procedural knowledge only through repetition.

When we asked the location of the "u" on the keyboard, your brain searched for concepts and images. When you are engaged in an action within the environment--for example, before the keyboard--the environmental cues and your actions pull one sequence of action after another.

Operationally we are pull systems. Designing pull (vs. push) systems is a key factor in the success of the lean strategies that have evolved from the Toyota production system.

The point is that eliciting VOC or mining the know-how of one organization and then implementing these findings into another organization involve translating tacit knowledge encoded procedurally in customers or other social spaces into explicit knowledge--expressing it symbolically.

That knowledge then has to be translated back into tacit knowledge. A lack of understanding of the difficulties in this process has doomed many an attempt to apply lessons learned from customers or another industry. Yet, strategies and tools to successfully and repeatedly do so are available.

Science based, time-tested tools such as QFD can be used to harvest wants and needs from customers and to translate technical know-how from one industry into another. VOC tools from QFD capture expressed and unexpressed wants of customers and identify underlying needs. Customers then group and rank the needs, and the organization can then focus its resources on highly ranked targets.

Five-Step Translation Process

There are also strategies, often called lead user strategies, to harvest the implicit know-how from other industries.¹⁰ Capturing either VOC or lead user know-how and translating that knowledge into actions of individual staff is at least a five-step process, illustrated in Figure 1 (p. 58):

1. Tacit knowledge, including VOC or know-how from another industry, is translated into an explicit form--as demanded qualities or organizational targets.
2. The explicit VOC is translated into an explicit voice of the organization--targets with measurable dimensions, or organizational performance measures.
3. Organizational performance measures are translated into targets and measures for key processes--process measures.
4. Process measures are translated into explicit targets and measurements for critical tasks by staff--task specifications.
5. Task specifications--know-what--must be embodied into day-to-day actions or automatic behaviors of staff that become new tacit know-how--or new habits.

In this sequence newly harvested explicit knowledge (step one) is combined with knowledge within the existing organization, steps two, three and four, and then reinternalized as tacit know-how that shows up in the day-to-day actions of people (step five).

Just as in the game of telephone, each step is an opportunity for a breakdown in translation, a loss of voice, introduction of inefficiencies and unnecessary cost, and a chance for adverse outcomes.

The whole process--from VOC or knowledge from another industry through the necessary repetition to encode new habits--is only a hypothesis about what is needed. Results must be monitored, and the hypothesis must be confirmed, rejected or modified by sampling the effectiveness of the results--relative to customer (or patient, in the case of healthcare institutions) concerns.

A Hospital Setting

Consider a simple service example from a hospital setting. Patient satisfaction data indicate patients and families are least satisfied with the timeliness of getting medication and with preparation for discharge back to the community. A process improvement team is charted to identify weaknesses in the process and improve results.

The team uses VOC tools and identifies faster relief of pain (not timeliness of getting medication) and getting more information about medications as two of the most highly weighted qualities demanded by customers.

The team then identifies two process measures for the demanded qualities--the number of people who reported they are satisfied with pain relief 30-45 minutes after receiving medication and the number of patients or family members who can identify, describe and report name, class, indication and major side effects of their medications prior to discharge.

The process is redesigned to test patients on their knowledge of medications just prior to discharge. Knowledge of diet and posthospital exercises are also to be tested.

A process owner or manager is identified. Staff members receive in-service training about the need to check and document patient satisfaction with pain relief after pain medication and the planned predischarge testing.

A review of charts a month after discharge (the most common quality assurance strategy in hospitals) notes only about half the charts document whether the patients were satisfied with pain relief after they received their medications and only 30% of patients or families passed the predischarge testing.

Consider for a moment the wisdom of retrospective chart review as a quality assurance strategy. What would you think of a vendor who shipped you a new computer system and did not check to see all components were in place before shipping but instead checked the packing list a month or two after shipment? What would you think of a computer manufacturer that checked the production documentation to see if a computer worked rather than turning the system on before shipping?

Testing equipment and checking products beforehand seem like common sense. But in hospital settings, most quality checks to ensure the patient receives safe and appropriate care are not carried out until one or two months after the patients have been discharged.

Furthermore, hospitals usually assess whether a treatment worked not by talking with or testing the patient but by reviewing the chart long after discharge.

Retrospective chart review is inefficient. It is also not very effective in changing procedural behavior. Unlike Pavlov's experiments, it is like ringing the bell a month after presenting food to the dog. The feedback is too late.

In-service training of staff members--showing them data or describing new targets and expectations--is an attempt to push knowledge into them. This approach ignores the fact that conceptual (explicit) knowledge and procedural (tacit) knowledge are distinctly separate systems in the brain, and encoding knowledge into procedural systems, such as learning to ride a bike, requires actual practice and repetition.¹¹

In this example, when the frontline nursing staff members were informed their patients were failing the predischarge test, they suggested if the organization was serious about improving the patients' capabilities, medication training would need to be started shortly after admission, not just before discharge, and would need to be repeated multiple times.

The project team responded to these initial defect rates (50% for documenting satisfaction/dissatisfaction with pain relief and 30% for patient knowledge of their medications) and to frontline staff suggestions by placing a prompt on the medication administration sheet on which nurses documented whether the patient was satisfied with his or her pain relief. Prompts were placed early after admission for the nursing staff to begin training patients and families on medications--a pull strategy.

Daily peer audits for satisfactory pain documentation completion were instituted and the results displayed on a visual feedback board on each unit--green (95-100%), yellow (90-94%) and red (90%). The completion rate of documentation of patient satisfaction with pain relief quickly rose to 100%, and over time the number of patients reporting satisfaction with their pain relief also increased.

Medication training was another story. Line managers were given the results of the predischarge testing process performance measures, and the pass rate for diet and home exercises improved. But patient knowledge of medication lagged.

The process manager verbally reported the poor results to nursing managers and used the organization's procedure for documenting nonconformance to specifications--which in this case was lack of medication training prior to discharge.

While the project team was trying to improve patient satisfaction with pain relief and patient safety with medications after discharge, the nursing leadership was busy trying to improve discrepancies between medication administration records and pharmacy charges to capture lost charges. Because the organization lacked a rigorous quality management system, this organizational disconnection was not addressed at the level of the organization as a whole.

Common Translation Breakdowns

Like this example, many change efforts fail to achieve their full intent because they focus primarily on the voice of the process or the tasks required by staff, while feedback necessary to create staff repetition is weak or missing (see Figure 1, p. 58).

The initial attempt at improvement in this example focused on communicating the need to change and the targets. But, in the case of the medications, no structural changes were made to prompt and hold staff accountable for new behaviors.

Failure to create in-process feedback that is direct and immediate at the staff level ignores human biology (step five). When direct, immediate feedback was introduced, documentation and patient satisfaction improved. In contrast, the same strategy was never fully deployed for patient medication training because the nursing leadership was focused elsewhere.

In this example, neither new nor existing organizational performance measures, which hold leadership accountable, were addressed as part of the improvement project.

Failure to identify, prioritize and rationally allocate resources and define accountability for overall organizational performance measures to hold leadership and the organization as a whole accountable is another common reason for project failure (step two).

Historically, hospitals arose as community based organizations with strong social values. In today's hospitals, however, financial measures routinely beat measures of social value when competing for resources.

The nursing leadership in the example was preoccupied responding to the need to improve a financial measure--capture lost revenue--while the hospitalwide organizational improvement effort focused on patient satisfaction and safety. Yet as noted in a recent article, one in five patients experiences an adverse event within days of discharge.¹² In elderly patients, two of the leading causes for readmission to the hospital after discharge are taking medication improperly and not following the proper diet.

Organizational Change

Organizational change efforts can be much more effective than they have been. The good news is that:

• The technical know-how to deal effectively with the breakdowns afflicting organizations already exists in other social spaces.
• The major barrier to incorporating that know-how is our own biology--our individual and collective inertia.
• Catalysts (quality tools and information technology processes) to overcome this inertia have also already evolved.

The path will require the application of science and the use of tools and IT to design processes that are customer focused for the delivery of services and management practices that are systems based.

In Consilience: The Unity of Knowledge, Edward O. Wilson describes science as "a two-step procedure by which natural scientists generally work top down across two to three levels of an organization at a time by analysis, then bottom up across the same levels by synthesis,"¹³ as illustrated in Figure 2 (p. 59).

For example, the successful treatment of Parkinson's disease required understanding the cause deep within the brain for the signs and symptoms in the body. Then, treatment strategies were designed to correct the deficit in brain function and thus also to correct the signs and symptoms in the body. Solving problems in organizations is no different.

The application of the principles of science to business has traditionally begun by studying consumer complaints about existing problems (defects) using an analytical approach in which you start at the downstream point (finished product) and move upstream.¹⁴

Analysis, however, can only go so far. It can reduce defects within existing products and services, but to go further or to introduce new features or products requires design. For new products or services, the approach must be from the upstream end looking downstream toward the qualities consumers demand in the finished product (see Figure 3).

The science of medicine has learned how to increase the understanding of disease by extending the level of analysis to the gene. To improve our understanding of defects, to effectively design new features into our production and services or to incorporate new knowledge from other industries, we need to extend our analysis of organizational processes and functions to the level of the human brain and to account for its capacities and limitations.

When we design fixes or deploy strategies inconsistent with how our brains function--using at least two distinct knowledge systems--the translation process breaks down and we complicate efforts at improvements.

REFERENCES

1. Yoji Akao, Quality Function Deployment: Integrating Customer Requirements Into Product Design, Productivity Press, 1988.

2. Dr. Edward Chaplin and John Terninko, Customer-Driven Healthcare: QFD for Process Improvement and Cost Reduction, ASQ Quality Press, 2000.

3. John Terninko, Alla Zusman and Boris Zlotin, Systematic Innovation: An Introduction to TRIZ, St. Lucie Press, 1998.

4. Larry R. Squire and Eric R. Kandel, Memory: From Mind to Molecules, Scientific America Library Series, 1999.

5. Kujiro Nonkaka and Hirotaka Takeucho, The Knowledge Creating Company, Oxford University Press, 1995.

6. Squire, Memory: From Mind to Molecules, see reference 4.

7. Nonkaka, The Knowledge Creating Company, see reference 5.

8. Yoji Akao, "QFD for the Twenty- First Century," keynote address at International Symposium on QFD, Novi, MI, 2000.

9. Chaplin, Customer-Driven Healthcare, see reference 2.

10. Mary Sonnack, lead user research tutorials, 12th Lead User Strategies Symposium on Quality Function Deployment, Novi, MI, 2000.

11. Squire, Memory: From Mind to Molecules, see reference 4.

12. A.J. Forster, H.J. Murff, J.F. Peterson, T.K. Gandhi and D.W. Bates, "The Incidence and Severity of Adverse Events Affecting Patients After Discharge From the Hospital," Annals of Internal Medicine, Vol. 138, pp. 161-167.

13. Edward Wilson, Consilience: The Unity of Knowledge, Alfred A. Knopf, 1998.

14. Akao, Quality Function Deployment, see reference 1.

EDWARD CHAPLIN, M.D., is medical director of Continental Rehabilitation Hospital, San Diego. He earned his doctorate in medicine at the University of Louisville. Chaplin is a member of ASQ and author of Customer-Driven Healthcare: QFD for Process Improvement and Cost Reduction, published in 2000 by ASQ Quality Press.

YOJI AKAO is a professor in the graduate school of business administration at Asahi University in Tokyo and has led or participated in research committees of the Union of Japanese Scientists and Engineers, the Japanese Standards Association and the Japanese Society for Quality Control. He earned a doctorate in quality control and statistical engineering at Tokyo Institute of Technology. Akao is a member of ASQ and received ASQ's Distinguished Service Medal in 2001.

Please Comment
If you would like to comment on this article, please post your remarks on the Quality Progress Discussion Board, or e-mail them to editor@asq.org