PERSPECTIVES

Drs. Poolman and Bhandari discussed healthcare research in orthopaedics. Our discussion uncovered 3 core themes. These themes and their supporting insights are highlighted below.

1. Scope Of Research

Dr. Poolman’s research focuses include identification of gaps in guidelines, quality improvement for patient care, the procedures that are being done most frequently and the conditions where the most money is involved. One of the topical questions under investigation is prognosis of patients after total knee replacements (TKRs) for treatment of osteoarthritis. About 20% of patients who had TKR still have complaints and Dr. Poolman is interested in further investigating the reasons. Another topic is cemented versus uncemented for primary hip replacement, using a regression discontinuity design to analyze registry data.

 

In registry studies, a lot of patient data are available and it is possible to include a complete population in the analysis. For instance, there are over 400,000 patients in the Dutch registry for implants which have covered 99% of all hip arthroplasties, including patients from all age groups. In orthopaedic practice in the Netherlands, surgeons tend to use cemented implants for older patients and uncemented implants in younger patients. The age cut-off points at different geographic locations in the country are different. For example, in Dr. Poolman’s hospital in Amsterdam, the age cut-off point for females was 65 years. Female patients older than 65 years with a higher prevalence of osteoporosis would receive a cemented hip in total hip arthroplasty (THA). Women younger than 65 years old would receive an uncemented hip. Dr. Poolman and the team are investigating the most optimal cut-off point for age that can be implemented across the Netherlands, using an analysis method from econometrics. Because randomized trials are not possible in the specialty of economics, economists use a regression discontinuity design to investigate causal inference in existing data. Dr. Poolman found that this regression discontinuity design was applicable in the orthopaedic registry studies. In order to find the best cut-off age, it would not be reasonable to compare the outcomes of a 50-year-old woman (who received uncemented hips most of the time in practice) and an 80-year-old woman (who received cemented hips). But groups of patients who were 64 years old and who were 66 years old are comparable. In the TKA registry database, there are about 20,000 patients whose age falls into the range of 64 and 66 years. Dr. Poolman regarded that with the econometrics’ regression discontinuity design, they were able to get some causal inferences to address their research question.

 

2. Values Of Real-world Evidence

In Dr. Poolman’s opinion, not only are the results from randomized trials important, but real-world evidence from natural experiments is also valuable. In the Netherlands, hospitals can choose either posterior or direct lateral approaches as their standardized treatment for hip fractures, hence practice in hospitals at different geographic locations is different. Dr. Poolman regards that it is meaningful to collect these data in a registry and to investigate any difference in patient outcomes. Because all patients are included, findings from such natural experiment studies have good applicability in clinical scenarios. Also, a stepped experimental design is going to be effective in enhancing implementation of a guideline and assessing its impact.

3. Involve All Groups in Research

Among the orthopaedic guidelines that Dr. Poolman and the team evaluated, about 50% of them were based on Level III or IV evidence. They are weak recommendations and actually are unanswered questions. The team then started to ask the guideline developers to also recommend future research questions. Various groups of representatives from guideline committees, the guideline working groups, the orthopaedic association, patients and insurance companies all participated in a process to prioritize the unanswered questions and to form a research agenda.

In Dr. Poolman’s value-based healthcare group, patients who underwent experimental treatment were involved. Dr. Poolman shared a typical example of patient involvement in research question generation. During a project of hip arthroplasty for osteoarthritis, two patients who were involved in the study questioned the investigators why they needed to sleep on their back following the surgery. Although the traditional practice was to instruct the patients after THAs to lay on their back and to use crutches for six weeks, one of the participating patients told the investigators that they were not comfortable laying on their back and wanted to lay on their side. Dr. Poolman and the team looked into the literature and found trials that investigated the patients’ question to see whether or not they needed to lay on their back. The evidence did not find that the position of laying on one’s side or having less restrictions resulted in more dislocations in patients who underwent THAs. The amended post-surgical care strategies were then implemented in a group of seven hospitals in the Netherlands. There was no increase in dislocation rate with the amended strategies.

 

Dr. Poolman and the team’s efforts at present, and in the next level are to involve not only orthopaedic surgeons and patients, but also other specialties including radiologists, rehabilitation doctors, infectious disease specialists and microbiologists in the development of modular research question agendas with a wider and a broader perspective in order to find real answers in a collaborative way.

SENSE-MAKING

Drs. Poolman and Bhandari discussed the value of real-world evidence in the form of natural experiments and studies using registry data. By including all the patients, real-world evidence has good applicability. In research, it is important to involve all groups of people, especially patients in research.