Six issues are addressed that complicate the use of current outcome measures to answer what many physicians, prosthetists, and therapists believe are the key questions facing amputation surgery and prosthetic rehabilitation. These issues complicate the isolation and measurement of the specific study variable, and therefore limit definitive conclusions that can be reached from both current and historic research trials. These six issues are:
1. A valid and reliable tool to document socket fit and socket comfort is currently not available.
2. The day-to-day variation in activity for lower limb amputees using prosthetic devices may well be larger than we acknowledge.
3. The accommodation time to a new prosthesis or to new prosthetic components has been poorly defined and measured.
4. The co-morbidities among the populations we are trying to study can have a large impact on function that often overwhelm the variable we are trying to study.
5. The use of multiple outcome measurement tools in attempts to address the issues of co-morbidities is a real phenomenon that can often lead to question overload, which can overwhelm many of our research participants.
6. Prosthetic outcomes research suffers because the complexity, the custom nature of the intervention, and limited availability of research subjects lead to small trail sizes, long trial duration, and very expensive trials.
1. Documenting socket fit and socket comfort is not standardized.
The majority of our outcome trials address the issues of standardizing an acceptable socket fit using the phrase “Socket fit and alignment were verified by a certified prosthetist.” While this has become the common solution to this difficult problem, it does not address the issue. A leading factor in lower limb prosthetics is socket fit and comfort. Poor or suboptimal socket fit can far outweigh the feet, shock absorbing pylons, knee units or other variables that we are trying to study.
We have no agreed upon, validated measurement tool to assess the fit and comfort of an individual's socket. A tool to measure socket fit and comfort would have immediate clinical and immediate research application. One of the highest priorities for outcome tool development would be for a validated tool to measure the success of socket fit and comfort.
2. The day-to-day variation in activity for lower limb amputees using prosthetic devices may be large and is often not measured.
There is much more day-to-day variation in functional activity, walking ability, walking symmetry, forces than we believe. This variation is larger than we estimate both in the first year, and also in established amputees. Figure 1 shows the data collected for one full year following one man's transtibial amputation for chronic pain and dysfunction. This man had a work related injury in 1991, and underwent multiple surgeries from 1991 to 1999. In 2000, at the age of 39, he requested transtibial amputation to decrease his pain and improve his ambulatory function. This man agreed to wear a step activity monitor on both his sound limb and on his prosthesis once he obtained it following surgery.
The chart shows tremendous day-to-day variation in activity, which persists far greater than expected during the post-operative period. In fact, the variation may not go away. The phenomenon of good days and bad days is a very real, and poorly defined aspect of life for an individual with limb loss. Patients indicate that this variation in comfort and walking ability does impact their activities of daily living, work activities and recreational activities. We have not done a good job of defining this variability.
This intrinsic variation makes precise measurement of function difficult because while the tool used to measure it might be accurate, the daily variation may well out-weigh the hypothesized effect. Measurements might well be made on a good day pre-intervention and a bad day post-intervention; or just the opposite. We do not really understand how many days of data collection are needed to fully understand the variation. The variation needs to be better defined and better measured.
3. Accommodation time for a new prosthesis or new prosthetic component is poorly understood.
Many individuals with limb loss discuss the fact that it took far longer to adjust to a new prosthesis than they had expected. We really do not know what the accommodation time is for sockets, liners, knees, and feet. Accommodation time may well be very different for different people trying to accommodate to the same new components or system. Again, look at the chart of 12 months of post-operative step activity data shown in Figure 1. There is a tremendous decrease in activity when this man obtains a new prosthesis. Two months after the new device was delivered, his activity is still far lower than it was when he had a poorly fitting system that needed to be replaced.
A single study has “quantified” accommodation time.1 This paper studied only one subject (age = 26, male, traumatic amputee from a motor vehicle accident), with a unilateral knee disarticulation. In this study, two knee units were interchanged: an Otto Bock 3R46 (Polycentric hinge, hydraulic swing control), and an Otto Bock 3R30 (polycentric hinge, friction swing control). The subject spent 4 weeks total test time on 3R46, and 25 weeks total test time on 3R30. The variables evaluated were: vertical ground reaction force, step and stride kinematics, preferred walking speed, vertical heel rise, knee angular velocity and time to reach peak vertical force.
The study concluded: While clinical decisions may be made after one week it is suggested that, for the purposes of research, an altered limb should be worn for at least 3 weeks to ensure that consistent walking performance is [to] be attained.
In a Cochrane Report from 2004, Hofstad et al.2 reviewed the literature regarding prosthetic components. Criterion B8 (Timing of Measurement) states: “This criterion pertained to the moment that the outcome was assessed in relation to the time subjects were given to adapt to a prosthetic change. An adequate adaptation period was required” (p 4). Also in 2004, van der Linde et al.3 expand upon this statement in the Journal of Rehabilitation Research and Development: “This criterion pertained to the moment that the outcome was assessed in relation to the time subjects were given to adapt to a prosthetic change. An adequate adaptation period was required. According to English et al.,1 transfemoral amputees need at least 3 weeks of walking with a new knee mechanism to be sure that gait parameters are stable. According to English's results and based on clinical experience, the amputees are assumed to need a period of at least 1 week to adapt to a new prosthetic foot or to a change in the prosthetic mass” (Appendix p 1–2).
4. Co-morbidities amongst the populations we are trying to study may have a large impact on function that overwhelms the variable we are attempting to study.
Individuals with diabetes and vascular disease have many co-morbidities that include: pulmonary disease, heart attack, congestive heart failure, strokes, transient ischemic attacks of the brain, renal failure, and large variations in fluid retention. The wide range of functional impairment these complex co-morbidities lead to can be huge and may well overwhelm any impact of prosthetic and orthotic intervention. The functional consequences that these medical co-morbidities have are also very likely to change during the research trial, often more than the variable we are trying to measure. In trauma, especially in those soldiers returning from recent wars, the single unilateral amputee without other limb fractures, lung, chest or abdominal injury, head injury, or psychological injury is very rare. These associated injuries have an impact on function that is difficult to measure. Although overall functional outcomes in these soldiers can be measured, the impact of different prosthetic devices, surgical techniques, or rehabilitation protocols will be quite complex to sort out. The typical solution to the management of a large number of co-morbidities is to conduct clinical trials with extremely large numbers of subjects so that the effect of the co-morbidities is diminished. Amputation and prosthetic trials with thousands of research subjects are simply not feasible.
5. The use of multiple outcome measurement tools in attempts to address the issues of co-morbidities is a real phenomenon that can often lead to question overload that overwhelms many of our research participants.
The concept of question overload in current amputation and prosthetic research and clinical trials is a real and growing issue. Many individuals in our studies lose focus and interest during the testing day when they are asked to complete the large number of forms included in the protocol. The forms and questionnaires all had good scientific rationale and were certainly included with good intent. However, it can be extremely difficult to sit through a session trying to answer multiple forms such as: SF36, PEQ, Brief Pain Index, depression form, sleep form, etc. How does “question overload” impact responses? While outcome tools for measuring specific disease states are necessary to address and understand the impact of co-morbidities, they also pose a problem that is in essence a “catch-22.” Measuring more factors often leads to overload, and possibly less effective measurement. This dilemma is not easy to solve, and a balance must be found.
6. Prosthetic outcomes research suffers because the complexity, the custom nature of the intervention, and limited availability of research subjects lead to small trail size, long trial duration, and very expensive trials.
Most drug trials or traditional orthopaedic implants are tested on animals and then tested in large trials that often include thousands of subjects. It is unlikely it will ever be possible to conduct a large trial because of the custom nature of our devices, and the associated costs and time commitments required. Fortunately for mankind, but unfortunately for science, there are simply not that many amputees who can be study subjects. Without the possibility of large trials, selection bias has a much larger impact on our results that we would like to acknowledge.
Human subjects issues put in place to protect individuals who participate in research trials, can also impact the outcomes of those trials. Regulations make it quite clear to individual subjects that they may withdraw from a study at any time. With high technology prosthetic component trials or even with newer surgical techniques on nerves and bone, our research subjects can and do stop participation on a whim. Some have discussed this phenomenon as the prosthetic equivalent of “dine and dash.” Some individuals seek to participate in research only as a mechanism to obtain high technology prosthetic devices that have been denied them by their health care insurance. Selection of subjects whom the research team believes are committed to the humanitarian side of research, and will not enroll only to seek personal benefit certainly can induce selection bias.
Higher technology trials, with longer accommodation time and measurement cycles means selecting subjects who are not likely to get sick, move, or have dramatic changes in their health or social situations. Again, this introduces a selection bias as researchers select only the more optimal subjects to participate.
These challenges are highlighted to foster discussion and communication on these issues and to stimulate new thinking and methods to address these challenges.