Effectiveness of Glenohumeral Arthroplasty

Last Updated: Thursday,  January 27 2005
We have implemented a system which we call "Carolina Codman which relies on self-assessment by the shoulder patient. It includes: self-assessment of function: The Simple Shoulder Test and self-assessment of health status: The SF 36. These data demonstrate the effectiveness of shoulder arthroplasty in defined circumstances. The following is a more formal paper describing some early results with this method.


Each physician needs to be able to demonstrate the effectiveness of the management programs he or she uses for different conditions. The challenge for the active practitioner is to evaluate this effectiveness in an efficient yet scientifically sound manner. The purpose of this study is to demonstrate a clinically practical method by which an individual physician can measure his or her personal effectiveness. The method used in this study is based on patient self-assessment.

Patients meeting strict criteria for primary glenohumeral degenerative joint disease completed standardized health status and shoulder function questionnaires before having a total glenohumeral arthroplasty by the author. The patients again completed the questionnaires at an average of ten months after surgery. Comparison of the pre and post surgery responses demonstrated highly significant improvement in the patients' assessment of their overall bodily pain and physical function as well as significant improvements in their role functioning and the anticipated change in their general health status. Concurrently these self-assessments indicated highly significant improvements in the patients' ability to sleep on their side to tuck in their shirt behind to place their hand behind their head to place various weights on a shelf at shoulder level and to toss overhand. They also indicated significant improvements in their ability to carry twenty pounds at their side to wash the back of their opposite shoulder and to do their usual work.

Click to enlarge
Table 1 - Health status score
Table 1 - Health status score

The goal of reconstructive orthopedic surgery is to improve patients' function and sense of well-being. It is important for the field of orthopedics to establish practical methods by which a surgeon can demonstrate his or her individual effectiveness in achieving this goal. Over the past decade standardized and validated self-assessment tools have become available which allow patients to characterize the quality of their lives and their physical function. Using these tools before and after treatment offers a practical method for demonstrating treatment effectiveness from the perspective of the patient.

The aim of the prospective investigation reported here was to use these self-assessment tools in characterizing the early effectiveness of an individual surgeon's total glenohumeral arthroplasty program in the management of patients with primary glenohumeral degenerative joint disease.

Materials and Methods

Since January 1992 all new shoulder patients seen by the author have been asked to complete two standardized self-assessment questionnaires: the SF 36 to define their pretreatment general health status and the Simple Shoulder Test (SST) to define their pretreatment shoulder function. The results from these questionnaires serve as the baseline for evaluating treatment effectiveness from the perspective of the patient.

The prospective study published here concerns all twenty-nine consecutive patients who both had a total glenohumeral arthroplasty by the author in the years 1992 and 1993 and met strict criteria for the diagnosis of primary glenohumeral degenerative joint disease. These criteria include (1) no prior history of trauma surgery or other known causes of secondary degenerative joint disease in the operated shoulder; (2) limited glenohumeral motion; and (3) radiographs showing joint space narrowing periarticular sclerosis periarticular osteophytes and absence of features indicating other causes of joint surface loss. The author's technique of total glenohumeral arthroplasty for primary glenohumeral degenerative joint disease has been described in detail previously. This series did not include patients with degenerative joint disease who were managed nonoperatively; thus there were no nonoperative controls.

Twenty of the patients were male and nine were female. The average age was sixty-five years (± thirteen SD) at the time of surgery. Sixteen of the patients had retired from work by the time of their procedure eight worked in less physical jobs (e.g. executive professor supervisor veterinarian) while five worked in more physical jobs (e.g. farmer electrician laborer).

Sixteen of the shoulders were on the patient's right side and thirteen on the left. Sixteen of the twenty-nine shoulders were in the dominant extremity.

In an attempt to make this type of research as cost-effective and simple as possible all followup SF 36 and SST questionnaires were sent to the homes of these patients at one time (first week of January 1994). All patients completed the followup forms on the first mailing and returned them at the latest by mid February 1994; no telephone contact or other prompting was necessary. The average time from arthroplasty to the followup self-assessment was 303 days ± 164 SD.

The pre and post operative SF 36 questionnaires were scored according to the system established by Ware et al. Pre and post operative SF 36 scores were compared using the Wilcoxon signed rank test to determine which of the general health status parameters were significantly changed after total glenohumeral arthroplasty. Pre and post operative SST results were compared using the paired rank test to determine which of the shoulder functions were significantly improved after surgery.

Correlation coefficients were also calculated among the improvements in each of the twenty one general health status and shoulder function parameters to determine the degree to which they were independent of each other.

To ascertain whether the changes in SF 36 and SST parameters were affected by the duration of followup the Mann Whitney U test was used to compare the results for the subset of fourteen patients having less than 300 days followup to those for the subset of fifteen patients having greater than 300 days followup after surgery. Correlation coefficients were also determined relating the change in each parameter to the duration of followup. Finally Spearman rank correlations were also carried out between the number of days after surgery and each of the health status and shoulder function parameters.

All data were entered by the author into a standard database (FileMaker Pro Claris) running on a laptop computer (Macintosh PowerBook Apple). The data were analyzed by the author on the laptop using standard statistical software (StatView Abacus).


At an average of ten months after total glenohumeral arthroplasty most of the general health status scores and shoulder functional assessments of these patients were substantially improved (See Table I). The Wilcoxon signed rank test revealed significant improvements in self-assessed overall bodily pain physical function physical role function emotional role function and anticipated health change. Similarly for ten of the twelve functions of the Simple Shoulder Test the paired rank test indicated that the percentage of patients stating they could perform the function after arthroplasty was significantly greater than the percentage who stated they could perform it preoperatively (See Table I).

Strong correlations did not exist among most of the twenty-one different parameters indicating that these assessments were relatively independent of each other. The highest correlation coefficients were observed for improvements in the following pairs of parameters: mental health and energy/fatigue (0.79) placing eight pounds on a shelf and washing the back of the opposite shoulder (0.59) placing a coin on a shelf and placing one pound on a shelf (0.54) physical function and energy/fatigue (0.54) tossing overhand and placing eight pounds on a shelf (.52) and placing a coin on the shelf and tucking in the shirt (.51). The highest correlation between an increment in a general health status parameter and a shoulder function was for physical role function and the ability to place the hand behind the head (0.51).

None of the improvements in health status or shoulder function parameters were significantly different between the two subsets of patients with different lengths of followup. The correlation coefficients between time of followup and each of the health status and functional parameters were all less than 0.2. The Spearman rank correlation rho values were all less than 0.5 except for the underhand throw which was 0.6.


Recently a great deal of interest has been directed to the determination of the overall effectiveness of surgical procedures including joint arthroplasty. However little attention has been focused on establishing the effectiveness of individual surgeons in the application of these procedures. It is axiomatic that each surgeon needs to know the degree to which his or her own surgery and postoperative management is effective in improving patients' health status and function. Practicing surgeons need an efficient way of collecting and presenting their personal effectiveness information to prospective patients and to payers of health care. It is insufficient for a surgeon to observe that a procedure is reported by others to be effective; rather it is necessary for the surgeon to demonstrate effectiveness in his or her own hands. This is a precept which should be credited to E. A. Codman a pioneering shoulder surgeon who presented the End Result Idea: "which was merely the common-sense notion that every hospital should follow every patient it treats long enough to determine whether or not the treatment has been successful and then to inquire if not why not? with a view to preventing similar failures in future." The study reported here is consistent with the End Result Idea except that the word "hospital" is replaced with the word "surgeon." In the final analysis the surgeon is the method.

Traditional approaches to clinical research require return visits radiographs physical therapy measurements and statisticians. These can all add up to an expense to patients and physicians that put such investigations out of reach in an individual surgeon's practice. Furthermore these "objective" measures do not assess what may be the most important determinant of a procedure's success: the patients' subjective evaluation of the improvement in their health status and function. Because self-assessment questionnaires can be completed by patients at home their use in measuring effectiveness optimizes the chances that all patients in a defined category will be included.

The SF 36 and SST questionnaires were well accepted by the patients in this study who completed them independently without difficulty or reluctance. Many were enthusiastic about being asked for their own assessment of their health status and shoulder function; some added letters with extended commentary regarding the benefits of the shoulder procedure on their lives. With a few exceptions post operative changes in the twenty-one parameters (the twelve SST questions and the nine health status scores of the SF 36) were independent of each other. This suggests that the number of questions could not be reduced without the loss of information.

In determining the value of a treatment to a group of patients it is critical to measure their status going into the treatment ("ingo") as well as their status coming out of it ("outcome"). In using standardized diagnostic criteria along with patient self-assessment of health status and function preoperatively the ingo to an operation can be characterized. By using the same self-assessment tools after surgery the outcome of the procedure can be characterized. The difference between the outcome and the ingo for the group treated by an individual surgeon is an indication of the effectiveness of that surgeon's application of the procedure. With this information the surgeon can communicate in patient-understandable terms both the usual preoperative status of individuals having the procedure and the expected results of the procedure in his or her hands.

The data presented here indicate that among a carefully defined group of patients with primary glenohumeral degenerative joint disease an individual surgeon's total glenohumeral arthroplasty program was effective within a relatively short time in improving both patients' shoulder function and general health status. The results were not significantly different for very short term followup (166 days ± 77 SD) and short term followup (431 days ± 108 SD). The long term benefits of total glenohumeral arthroplasty for primary glenohumeral degenerative joint disease remain to be documented using these tools.

This is one of the first studies to demonstrate statistically significant changes in a standardized health status self-assessment after shoulder reconstruction. More importantly it suggests a practical method by which a surgeon can demonstrate his or her individual effectiveness.

This study was accomplished without the expense or inconvenience of return visits research assistants or specialized computer software. Thus the method and the tools are practical and generic; they can be extended without difficulty to the measurement of the individual effectiveness of other operative and nonoperative management programs.