USING TIME-ORDERED DATA TO MAKE DECISIONS ABOUT PATENT CARE- PART 1 of 10

Providing quality care to patients requires analysing process data to make informed decisions. However, making the right decisions based on data is often complicated by natural variation in clinical and managerial processes—that is, if I take serial measures of anything such as BP, blood sugar, post-surgical complications, or patient waiting times, the subsequent number will always be better or worse than the previous one but this doesn’t indicate improvement or worsening situation.

Although traditional statistical analysis methods —mean, median, SD, T-Test, Chi-Square, F-Test— account for natural variation (P value <0.05), they require aggregation of measurements over time, which can delay decision making.

Conversely, the use of time-ordered data, aka Statistical Process Control methodologies—include run charts, shewhart control charts and CUSUM charts—is an intuitive, practical and robust approach to monitor and improve the quality of healthcare and deal with variation in clinical and managerial processes.

Pionered by physicist and engineer Walter Shewhart in the 1920s and later adopted to healthcare, SPC offers a framework to learn from variation in clinical and organisational processes.

Clinical variation from evidence-informed best practice can result in error, harm or poor patient outcomes while variation from organisational systems and processes leads to inefficiencies, waste of resources and increased waiting times.

For example, the run chart below shows daily systolic BP readings around a median line for a patient with hypertension. Question—how can I tell whether the intervention (diet, exercise or medication) is working by just looking at the data points?

SPC methodologies are now used extensively to monitor clinical outcomes (asthma, diabetes, hypertension etc ), surgical outcomes (inpatient death, unplanned admission to intensive care, reoperation, and a combination of severe complications—cardiac arrest, pulmonary embolism, sepsis, or surgical site infection), public health surveillance, and the learning curve of trainees undertaking medical or surgical procedures.

In part 2, I’ll be discussing the difference between common-cause and special-cause variation in statistical process control methodologies.

Dr Subiri Obwogo is a medical doctor, specialist in Public Health Medicine and Quality Improvement Advisor. [email protected]