“I think we can do better”. This quote from an Australian comedian is, in six words, the basis for quality assurance and, its close cousin, continuous quality improvement. There are a variety of different elements in the process, beginning with measurement and manipulation. The process requires reliability and validity. Constructs are examined and operationalized. Data are collected and analyzed. Hypotheses are tested. Theories are developed and supported, modified, or abandoned. Does this sound like scientific method? You betcha.

A variety of businesses have benefited from a quality assurance process. This relies on standards of performance, whether in manufacturing screws to exacting specifications or inserting those screws into luxury automobiles with precise torque. An acceptable range of performance is defined. When the range is not met for a certain predetermined percentage of products, an attempt is made to define the problem and propose a solution. Quality assurance is critical when the product has safety implications. If the screw holds the brakes in place, failing to meet the acceptable range of performance can have disastrous consequences. Therefore, the percentage of products meeting specifications must be extremely high, if not 100%. On the contrary, a burnt potato chip may irritate some consumers, but is unlikely to have serious consequences.

Quality improvement does not have an end goal and works in a more continuous fashion. When used in customer service, surveys may be used to obtain satisfaction scores. No matter what the score, an intervention is proposed to try and improve it. When the score is at the top of the range, a new measure of satisfaction is developed that leaves room for improvement.

Quality assurance and improvement processes are both important in medicine, specifically in sleep medicine. There are some areas where the goal should be a total absence of events. For example, patient deaths in the sleep center should not be merely infrequent, they should never happen. These “sentinel events” should result in an immediate change in policies to ensure that they do not happen again. Improving care in sleep medicine is an imperative that benefits sleep centers and patients. The goals are to improve the quality of care, improve the health of populations, and reduce the costs of health care (1).

Measures can be divided into structural, process, and outcomes (1). For example, if you want to know if the number of patients visiting your sleep center is higher today than it was a year ago, you can review billing data or procedure logs. This is considered a structural measure. If you want to know if visiting the sleep center results in reduced blood pressure on the next clinic visit, you can review vital signs from the patient’s chart. This would be a process measure. It is nice to have blood pressure in the normal range, but in itself this conveys no specific benefit. The benefit to the patient would be a reduction in the risk of stroke or heart attack, and the number of these events is considered an outcome measure. Outcome measures often require an
extended period of time to collect, and as a result may suffer from lower reliability.

The first step in quality assurance or improvement is to measure. The first question is, “What am I interested in?” and the second question is, “How can I measure it?” The first question asks for a construct, and the second asks for an operational definition. This can range from a simple question with obvious measures to more complex constructs that require more complicated measurement tools.

Here’s an example from home sleep apnea testing (HSAT) that starts with a simple question and ends with a more complex question. The simple question is, “How many studies do I need to repeat?” The simple measure is to count the number of studies that were considered inadequate and required retesting and divide by the total number of studies run per month. Setting a goal for a small percentage of inadequate studies has implications for patient satisfaction and profit margin. Patients do not want to be told that they need to repeat a study. This delays determination of the results of the study and requires additional work, causing discomfort to the patient. In addition, current reimbursement rates provide a small margin between the cost of the HSAT and the revenue produced. A few inadequate studies per week might make HSAT unprofitable, leading to discontinuation of the service and an increase in patients who are not diagnosed. A high percentage of inadequate tests tells you there is a problem but gives no insight into how to correct the problem. In our fictional sleep center, the Galileo Sleep Measurement Center, more than 20% of our HSATs are rejected by the physicians. Something must be done. But what?

The more complex question is, “Why were the studies considered inadequate?” The idea of an inadequate study is a construct that needs to be defined operationally. That is, we need to make it measurable. This typically starts with observation. We could put a box on the interpretation form that asks the physician, “Why was this study inadequate?” These open-ended questions are difficult to quantify because similar responses can be provided with slightly different language, making it difficult to decide if the causes are actually the same. One physician writes, “oxygen problem” and another physician writes “saturation measurement lost.” Can we offer the physicians some options with check boxes? The American Academy of Sleep Medicine (AASM) Scoring Manual (2) provides a starting point. The features of an acceptable HSAT device must be met in order to provide an adequate study. These features include oxygen saturation and heart rate, as well as a measure that allows for the calculation of a surrogate of the apnea-hypopnea index (typically airflow for the numerator and recording time for the denominator). Of the 20% of recordings rejected by the physicians in our fictional center, 75% are identified as oxygen saturation problems.