Abstract
Background:
The aim of this study was to determine whether immobilization of an arm has detrimental effects on driving performance.
Methods:
Thirty-six healthy officers-in-training were assigned a sequence of fiberglass splints (left and right-sided above-the-elbow thumb spica and below-the-elbow splints) with use of a randomized higher-order crossover design. Runs were scored on a cone-marked driving course used for officer certification with predetermined passing requirements. Driving time, the number of cones hit per course section, and the cone-adjusted total time (a five-second penalty per hit cone) were recorded. A linear mixed-effect model with random environmental and learning effects for cone-adjusted time analysis was used. Participants rated perceived driving difficulty and safety with each splint, and ratings were compared with the Wilcoxon signed-rank test.
Results:
Thirty participants completed the entire set of runs. Analysis of total cone-adjusted time revealed a significant performance decrease with the left arm in an above-the-elbow thumb spica splint (average, 22.2 seconds; p < 0.001) and with the left arm in a below-the-elbow splint (average, 16.2; p = 0.007). Analysis of forward-only course sections revealed poorer performance trends with all splints, with the worst performance with the left arm in an above-the-elbow thumb spica splint. Driving with the left arm in an above-the-elbow thumb spica splint had the highest perceived difficulty (median, 8.0) and lowest perceived safety (median, 3.0).
Conclusions:
Driving performance as measured with a standardized track and scoring system was significantly degraded with splint immobilization of the left arm. Further studies are required to determine the effect of arm immobilization on normal driving conditions.
The principle of extremity immobilization is a long-standing fundamental of orthopaedic care. Immobilization is currently used in the management of trauma, infection, and inflammation, with duration of treatment lasting from days to months. Therefore, the use of splints, casts, and other devices for immobilization of extremities is commonplace.
Surveys have shown that a substantial number of patients drive while wearing one of these devices1-3. A driver with an immobilized arm, however, may pose a health risk to the general population, and the treating physician may be exposed to medicolegal claims. In short, the decision of whether to let a patient with an immobilized arm drive can have personal, financial, societal, and medicolegal ramifications.
Unfortunately, data on which to base such a decision are lacking. It is not surprising, therefore, that physician opinion appears to vary widely with regard to when a patient should be allowed to return to driving1. This variation may be especially pronounced with immobilization of the arm4.
This study sought to provide evidence-based guidance for physicians to appropriately counsel their patients with regard to driving with an immobilized arm, on the basis of driving performance with above-the-elbow and below-the-elbow immobilization on a standardized closed driving course. Our primary hypothesis was that above-the-elbow immobilization, but not below-the-elbow immobilization, substantially worsens driving performance. Our secondary hypothesis was that driver performance is worse when his or her dominant extremity is immobilized in a splint.
We performed a randomized higher-order crossover trial to investigate the effects of arm immobilization on driving performance.
Patients and Protocol
Participants were selected from current officers-in-training who met the inclusion-exclusion criteria. Inclusion criteria were an age of eighteen years or older, full range of motion of the upper extremity without deficit, and adherence to standard proper driving technique as specifically instructed by the state Law Enforcement Training Academy. Exclusion criteria were a recent extremity injury affecting or limiting the ability to drive, current immobilization outside the study, or any medical condition precluding driving safely. All participants gave informed consent to participate in this study, which was approved by our institutional review board.
The study was conducted on the grounds of the state Law Enforcement Training Academy, which houses a standardized driving performance course (see Appendix) used for training of regional emergency personnel. All officers undergoing basic training in the region are required to pass this course. Designed to simulate the driving environment experienced by police officers while operating in emergency and nonemergency modes within a limited physical space, the course requires a variety of maneuvers that the officer in the line of duty may be called on to execute in an efficient, safe manner. Specific course tasks include braking, backing up, steering, cornering, acceleration, parking, and off-road recovery.
The trial was conducted over a single day, which was divided into five time periods, two in the morning and three in the afternoon. During a single time period, each driver completed a run through the course. At each time period, there were six drivers with each type of splint: control (no splint), left above-the-elbow thumb spica splint, right above-the-elbow thumb spica splint, left below-the-elbow splint, and right below-the-elbow splint. Drivers were randomly assigned a sequence of the five splints. Each driver also had a control run from a previous day of their training, for a total of six runs per driver.
Measurements of Driving Performance
Each run through the driving course was scored by experienced Academy driving instructors using the standard scoring protocol. Driving time and the number of cones knocked down are recorded at the end of a run, as well as at a number of points throughout the driving course. The cone-adjusted time is defined as the total time with five seconds added for each cone knocked over. According to the Academy's protocol, a run is considered a failure if the cone-adjusted time is greater than three minutes and fifty seconds, or the number of cones knocked down is greater than three. More details about the course and the score sheet can be found in the Appendix.
At the end of the day, the drivers were asked to complete a survey that asked them to rate both perceived difficulty and safety of driving with each splint.
Statistical Analysis
Cone-adjusted driving times with no splint were compared with cone-adjusted times with each of the four types of splints and the control from a previous day with use of a linear mixed-effect model that properly adjusted for the correlation of the six course times within a driver. The estimates of cone-adjusted time, along with 95% confidence intervals and p values, are reported. Medians and 95% bootstrap confidence intervals from 1000 bootstrap samples are reported for the survey responses with regard to perceived safety and difficulty, which are compared between cast types with the Wilcoxon signed-rank test.
For the two time periods in the morning, it was rainy and the course was wet; by the time lunch was over, it was sunny and the track had dried. Details with regard to how multiple imputation analysis accounted for this problem are available in the Appendix. Summary statistics, graphics, and linear models were generated with use of R statistical software (version 2.9)5.
Source of Funding
No external funding source was used in this study. Splinting material used in this study was donated by 3M (St. Paul, Minnesota).
Demographic Data
Of the thirty-six drivers who started the study, only thirty completed all five time periods on the day of the experiment. A block of six drivers was excluded after the first time period because of time constraints. The ages of the drivers ranged from twenty-one to forty-seven years (median, twenty-seven years). The group of thirty drivers included four women and eleven college graduates. Twenty-eight were right-handed, and two had eighty hours of training in emergency vehicle operation. Their driving experience ranged from five to thirty-one years (median, thirteen years).
Driving Performance
The drivers took an estimated 200.4 seconds (95% confidence interval, 192.0 to 208.8 seconds) to complete the course with no splint on the day of the experiment. There was no evidence of a difference between the time drivers took to complete the course without a splint on the day of the experiment and the time it had taken without a splint on a previous day (95% confidence interval, —18.8 to 0.3 seconds; p = 0.06).
The immobilization of the left arm in an above-the-elbow thumb spica splint slowed the drivers by an estimated 22.2 seconds (95% confidence interval, 10.0 to 34.4 seconds; p < 0.001); there was not sufficient evidence to show that an above-the-elbow thumb spica splint on the right arm caused any slowing (estimated slowing, 8.2 seconds; 95% confidence interval, —3.5 to 19.9 seconds; p = 0.17). Similar to the results with the above-the-elbow thumb spica splint on the left arm, the below-the-elbow splint on the left side slowed the drivers by an estimated 16.2 seconds (95% confidence interval, 4.6 to 27.7 seconds; p = 0.007), while the below-the-elbow splint on the right arm slowed the drivers an estimated 5.1 seconds (95% confidence interval, —6.7 to 16.9 seconds; p = 0.40), which was not significant (Table I).
The analysis above was repeated for the time taken to execute only the sections of the course that had not involved backing up the vehicle. On the average, these forward-only tasks took 86.9 seconds (95% confidence interval, 82.4 to 91.4 seconds) to complete with no splint. There was not sufficient evidence to conclude any of the splints caused slowing in the forward-only tasks. The above-the-elbow thumb spica splint on the left arm slowed drivers by an estimated 4.5 seconds (95% confidence interval, —1.5 to 10.4 seconds; p = 0.217), and the below-the-elbow splint on the left arm slowed drivers 3.8 seconds (95% confidence interval, —2.8 to 10.4 seconds; p = 0.260). The splints on the right side did not slow drivers as much as the splints on the left side, although the difference was not significant. The above-the-elbow thumb spica splint on the right arm slowed drivers 3.2 seconds (95% confidence interval, —3.3 to 9.7 seconds; p = 0.144) and the below-the-elbow splint on the right arm slowed drivers 2.1 seconds (95% confidence interval, —4.8 to 8.9 seconds; p = 0.558).
Perceived Difficulty and Safety
Drivers were asked to rate safety and difficulty on a scale from 0 to 10, with 10 indicating, respectively, the safest or the most difficult (Figs. 1 and 2). Perceived difficulty for a below-the-elbow splint did not differ between sides (median ratings, 2.0 for the right side and 2.5 for the left side; p = 0.08). Perceived difficulty differed between the sides for the above-the-elbow splints (median, 4.5 for the right side and 8.0 for the left; p < 0.001), as well as between the above-the-elbow and below-the-elbow splints on the left side (median, 8.0 and 2.5, respectively; p < 0.001). The above-the-elbow thumb spica splint on the left arm had the highest perceived difficulty, with a median rating of 8.0 (95% bootstrapped confidence interval, 7.0 to 8.0).
Survey data for perceived safety followed similar patterns (Fig. 2). Drivers felt safest with below-the-elbow splints (median, 9.0 for the right side and 8.0 for the left), and below-the-elbow splints on the right arm were perceived to be safer than below-the-elbow splints on the left arm (p = 0.02). An above-the-elbow thumb spica splint on the right was perceived as less safe than a below-the-elbow splint on the right (median, 6.0 and 9.0, respectively; p = 0.001). The above-the-elbow thumb spica splint on the left arm was perceived as the least safe, with a median rating of 3.0 compared with a median of 6.0 for the above-the-elbow thumb spica splint on the right arm (p = 0.001) (Table II).
There is likely a high worldwide prevalence of driving with an immobilized extremity. According to an Australian study of 168 sequential patients of driving age seen in a fracture clinic with an upper-extremity cast, 50% drove at least once and 22% drove daily while wearing the cast2. An Irish survey of 118 consecutive orthopaedic outpatient adults with either wrist (61%) or ankle fractures revealed that 15% drove with a cast3. A recent survey of seventy patients in an urban New England orthopaedic clinic found that 42% considered the inability to drive as a major difficulty, representing in itself a major financial hardship for 26%. Of those patients, 35% were driving despite narcotic use and 19% had a sense of danger while doing so1. Degradation in driving ability from an immobilized extremity could have considerable effects not only for the patient but also for the community at large. Scant prior evidence, however, is available to make an informed decision with regard to allowing patients to drive when they have an immobilized extremity, especially an upper extremity.
We found all forms of immobilization showed a trend toward worse driving performance by our measure compared with no immobilization. The splint on the left arm, especially the above-the-elbow thumb spica splint, was associated with significant performance degradation. Because of the size and demographic makeup of our study, we were unable to make a statistical analysis of the effect of hand dominance; nevertheless, the side of immobilization appears to have a significant effect on performance degradation. We theorize that our finding of worsened performance with a splint on the left arm may be due to the visual and spatial constraints inherent with a left-sided driver seat. With minimal space to the left of the driver, the immobilized arm may be forced into a position in the way of maneuvering the steering wheel. The splint on the left arm may be especially cumbersome when backing a vehicle; the driver must look over the right shoulder for adequate visualization, and he or she often places the right hand on the front passenger seat for stability, leaving the immobilized arm to maneuver the steering wheel. Post hoc analysis of this study suggests that backing up may play a large role in the significantly worse performance with immobilization of the left arm. Of the estimated 22.2 seconds of slowing that was observed with the above-the-elbow thumb spica splint on the left arm, only 4.5 seconds were attributed to forward driving. Our survey results correspond with our driving course findings. The above-the-elbow thumb spica splint on the left arm was clearly perceived as causing the most difficulty and as being the least safe of all of the tested splint types.
The current study is limited in several important aspects. The officers-in-training who participated in this study are a specialized group; thus, caution must be used in extrapolating these results to the general population. The average age of the participants in this study is likely younger than the average age of the general driving population. These young adult drivers with specialized driver training probably represent a subset of drivers that arguably performs better than the average driver. As such, one might expect that any negative effect of immobilization on these study participants would be similar or increased in the general population. These participants also were without extremity fracture, pain, or other abnormality. Any degradation in driving performance seen with immobilization would likely be worsened by the presence of such underlying or related issues. Only one automobile make and model was used for the current study: a four-door sedan with automatic transmission. It is unclear how other types of vehicles would affect performance. For instance, the increased complexity of driving with an immobilized extremity in a vehicle with manual transmission might be expected to further compromise performance. If the decreased driving performance seen when a splint was worn on the left arm was due in part to the minimal space for the immobilized left arm in a car with the driver's seat on the left side, this effect might be magnified in a smaller car. Some cars may have a larger blind spot, making looking over the shoulder when backing up even more crucial for safety. Because of the health and skill of the drivers and the type of car used, the results of this study may serve to provide a so-called best-case scenario of driving performance with an immobilized arm, given the testing parameters and the use of a closed course.
While the closed course was designed to replicate real-world circumstances, aspects of actual street driving may be different, and these aspects may affect driving performance with an immobilized limb. Although the course is designed to be quite complete, certain important driving maneuvers may not be tested, while other maneuvers, such as backing up or three-point turns, may be overrepresented in the course. Some skills particularly useful for emergency personnel, such as off-road recovery and backing out of an offset alley, may not be typical in regular street driving. It is possible that the effects of arm immobilization may be amplified with use of these more difficult maneuvers, although this was not specifically tested. Also, this course did not involve moving hazards or other random events that might occur in actual street driving and may cause a worsened outcome with upper-extremity immobilization, as was found in a simulator study6.
The scoring system, which rewards for driver efficiency and accuracy, may not be a precise marker of actual street driving performance. The timed aspect of the course may have led to artificial stress that would not have been present during normal driving. Drivers with an immobilized arm may have altered their driving practice apart from a timed study, such as slowing down. It must be noted that the officers-in-training understood that their driving performance during the study did not affect passage of their basic training course. Additionally, any additional stress from timing would have likely affected the control run of each driver in a similar manner.
The present study measured driving performance, which may not be an absolute indication of driving safety. The ability to skillfully perform on the road likely contributes to safe navigation; splints with the worst performance in this study also had the lowest perceived safety. However, there are many important considerations to ensuring safe driving, factors which are beyond the scope of this study.
Even with its limitations, the current study contributes substantially to the existing body of evidence with regard to driving with an immobilized arm. Lower-extremity function for driving has been commonly measured with brake reaction time7-11. Unfortunately, because of the complex nature of upper-extremity use in driving, no similar marker of driving function for the upper extremity exists. Still, simple attempts have been made to quantify the effect of upper-extremity immobilization.
In a British study with use of a driving simulator, eight healthy young adults performed four twenty-minute simulated driving circuits involving urban and rural conditions along with specific preprogrammed hazards6. The first and last runs were performed without immobilization; the second and third runs were randomized to either left-then-right or right-then-left wrist immobilization with a below-the-elbow cast. The authors concluded that participants were more cautious when the wrist was immobilized and showed decreased performance in response to hazards. Post hoc analysis revealed more pronounced deterioration with casts on the right side, which was believed possibly to be due to the right-hand dominance of all participants. Given our results, an alternate explanation is that more pronounced deterioration of driving performance occurs when the immobilized arm is on the same side as the driver's seat.
In a British single-subject study12, one of the authors serially wore three types of upper-extremity casts (Colles, scaphoid, and Bennett) and tested his ability to change gears, steer, reverse, hand-brake, control the indicator and horn, and drive around town. A self-scoring 3-point scale rated his "limitation of control." The author-driver concluded that while the Colles cast was not particularly limiting, the scaphoid and Bennett casts substantially impaired steering-wheel handling and hand-brake clearance.
Although studies such as those described have made simple attempts to determine the effect of immobilization on driving performance, the evidence of effect has remained quite limited. Thus, existing guidelines for driving restrictions are only able to provide general principles. The United States Public Health Service has suggested that, in order to drive, patients should have normal motor function of both upper extremities and the right lower extremity, along with "adequate" mobility of the joints13. The American Medical Association Council on Ethical and Judicial Affairs concluded that, for a physician to recommend against driving, the driver must pose a clear risk to public safety, and the physician must be able to identify and document impairments that clearly relate to the ability to drive14. The American Medical Association suggests no driving restriction for limb fractures and treatment involving splints and casts, provided there is no interference with driving tasks, although the target for this guideline is older patients15.
In keeping with the ambiguous guidelines for driving restrictions, physician agreement is lacking. A recent anonymous survey of forty-one New England orthopaedic surgeons revealed substantial variation in the criteria used to allow patients to return to driving after eight types of upper and lower-extremity fractures1. Sixty-six practicing orthopaedists in the United Kingdom were asked to determine whether they would allow driving in scenarios involving different fractures, treatments, and stages of fracture-healing. A majority of the physicians agreed on most lower-limb fracture scenarios but had poor agreement on upper-limb fracture scenarios4.
Given the lack of clear evidence-based recommendations for driving with extremity immobilization, our study makes an important contribution to the field. Compared with previous studies, the current study benefited from the use of a preexisting standardized driving course and scoring system for real-world certification. The driving environment used in the current study allowed for real vehicle maneuvering in the type of controlled environment required for rigorous statistical study. While participants and observers were not blinded to the type or side of immobilization, this was unavoidable in both the present and previous studies. Further studies are required to determine the effect of arm immobilization on driving safety. The current study is the first randomized controlled trial, as far as we know, to examine the effect of arm immobilization on driving performance, and the results showed that all forms of immobilization trend toward worse driving performance compared with no immobilization.
Figures showing the course layout and the rain and learning effects as well as a description of the course and details on the statistical analysis are available with the electronic version of this article on our web site at jbjs.org (go to the article citation and click on "Supporting Data").
Note: The authors thank Shana Roberts and the other members of the Tennessee Law Enforcement Training Academy for their invaluable partnership as well as Julie Daniels and Vickey Smotherman for their invaluable assistance in this study.
Chen
V;
Chacko
AT;
Costello
FV;
Desrosiers
N;
Appleton
P;
Rodriguez
EK. Driving after musculoskeletal injury. Addressing patient and surgeon concerns in an urban orthopaedic practice. J Bone Joint Surg Am.
2008;90:2791-7.[PubMed][CrossRef]
Kalamaras
MA;
Rando
A;
Pitchford
DG. Driving plastered: who does it, is it safe and what to tell patients. ANZ J Surg.
2006;76:439-41.[PubMed]
Kennedy
MT;
Roche
S;
Lenehan
B;
Curtin
W. Driving plastered: driving habits of orthopaedic outpatients and the medico-legal implications. Eur J Orthop Surg Traumatol.
2006;16:228-30.
Rees
JL;
Sharp
RJ. Safety to drive after common limb fractures. Injury.
2002;33:51-4.[PubMed]
R Development Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2009.
Gregory
JJ;
Stephens
AN;
Steele
NA;
Groeger
JA. Effects of upper-limb immobilisation on driving safety. Injury.
2009;40:253-6.[PubMed]
Egol
KA;
Sheikhazadeh
A;
Mogatederi
S;
Barnett
A;
Koval
KJ. Lower-extremity function for driving an automobile after operative treatment of ankle fracture. J Bone Joint Surg Am.
2003;85:1185-9.[PubMed]
Nguyen
T;
Hau
R;
Bartlett
J. Driving reaction time before and after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc.
2000;8:226-30.[PubMed]
Spalding
TJ;
Kiss
J;
Kyberd
P;
Turner-Smith
A;
Simpson
AH. Driver reaction times after total knee replacement. J Bone Joint Surg Br.
1994;76:754-6.[PubMed]
MacDonald
W;
Owen
JW. The effect of total hip replacement on driving reactions. J Bone Joint Surg Br.
1988;70:202-5.[PubMed]
Ganz
SB;
Levin
AZ;
Peterson
MG;
Ranawat
CS. Improvement in driving reaction time after total hip arthroplasty. Clin Orthop Relat Res.
2003;413:192-200.[PubMed]
Blair
S;
Chaudhri
O;
Gregori
A. Doctor, can I drive with this plaster? An evidence based response. Injury.
2002;33:55-6.[PubMed]
Stock
MS;
Light
WO;
Douglass
JM;
Burg
FD. Licensing the driver with musculoskeletal difficulty. J Bone Joint Surg Am.
1970;52:343-6.[PubMed]
CEJA Report I—I-99. Impaired drivers and their physicians. .
American Medical Association with the National Highway Traffic Safety Administration. Physician's guide to assessing and counseling older drivers. .