Methods

The Distal Radius Study Group (DRSG) was formed in 2003. Original group membership included four academic centers, one single-specialty group practice with academic affiliation, and one single-specialty group with no academic experience. The Primary Investigator (PI) acted in an oversight capacity and did not enroll patients.

The group was formed with the intent to examine the possibility of an Internet-based data set as a means of detecting differences in distal radius care. The data to be collected had been agreed on by group members at the outset and were presented from the fixed data set via the Internet to document enrollment, baseline data, treatment data, and follow-up data at specified intervals. The data set was kept constant throughout the study interval (December 2003 to February 2007). Two of the six sites never entered patients (the site housing the oversight PI and the single-specialty group with no academic experience). In 2006, a seventh site (fifth active) joined the DRSG and has successfully enrolled patients using the complete functionality of the collection system (real-time use with direct data transfer into an electronic medical record).

The study was designed in the manner of a PCT. As such, the inclusion and exclusion criteria were kept as broad as possible. The primary exclusion criteria were (1) open epiphysis, (2) open fracture, (3) history of inflammatory arthritic process, (4) fractures with associated nerve or tendon injuries requiring surgical repair, and (5) inability to read or write English. For the DRSG, institutional review board approval has been obtained and maintained at all sites.

The purpose of our choosing an observational model for the study was threefold:

Previously reported variables were used as a guideline to create structured data sets for demographics, baseline, operative (pathology, equipment, complications), outcomes, and adverse events. The Disabilities of the Arm, Shoulder, and Hand (DASH) score was selected as the patient-reported instrument and designated as primary outcome for comparison. Fracture type as designated by the Orthopaedic Trauma Association classification system for distal radius fractures was selected as the primary means of fracture/disease classification.¹⁰ A conceptual model for treatment of distal radius fractures was designed (Fig. 4-1). The model depicts the basic treatment/measurement process and identifies the primary outcome and independent variables.

FIGURE 4-1 Outcome Study Design Diagram for Distal Radius Fracture Treatment. The primary outcome is identified as the patient’s DASH score. Independent variables (fracture type, physician, treatment site, treatment type [device and technique], and patient) also are shown.

A study booklet was created for each site that summarized the data entry process and was used to familiarize the users with structure of the “screen” that would be seen during the process of data entry. This included a view of all patients enrolled and their follow-up schedule (divided into five postoperative intervals: 2 weeks, 6 weeks, 13 weeks, 26 weeks, and 52 weeks [±20%]) (Fig. 4-2). The type of data element to be collected was seen by the user and enforced by the system (range with specific end point using slider bars, single possible choice of data element from a list, multiple possible choice data element or elements from a list, and open text entry) (Fig. 4-3).

FIGURE 4-2 Screen Shot Showing Follow-up Schedule.

FIGURE 4-3 Screen Shot Showing Data-Entry Screen as Seen via the Internet.

To enable individual treatment site data entry, an Internet-based data entry method was designed. BoundaryMedical Incorporated (www.boundarymedical.com) developed, deployed, and hosted the method as employed. No software applications or databases were installed at any remote site computers. A monitor, Internet connection, and browser (any) were the only required products at the remote site. The methods met Health Insurance Portability and Accountability Act (HIPAA) compliance and data integrity and security standards.

Each site monitored completion of individual patient follow-up data of enrolled patients using a site-specific follow-up screen that was continuously updated and available via the Internet. In some cases, outstanding DASH scores were obtained via an Internet-based patient portal from home or elsewhere. Data were obtained for analyses using a secure download method into a specifically formatted Access database that was used to create relational comparisons that were statistically analyzed using SAAS.

Results

Since inception, 280 patients have been identified as meeting the enrollment criteria for study inclusion. Of these, 220 met all inclusion and exclusion criteria and chose to enroll. The primary author (M.D.P.) enrolled one third of all patients treated (Fig. 4-4). Enrollment was offered to qualified patients at the time of treatment initiation. Figure 4-5 depicts the distribution of patient ages, gender, and season at time of fracture. These and other patient factors (e.g., osteoporosis) appear similar across sites.

FIGURE 4-4 Enrollment Screen Showing Enrollment Criteria and Case Count.

FIGURE 4-5 Demographics Summary Sorted into Gender, Age, and Season at Injury.

Sites were instructed to offer enrollment to all patients seen regardless of planned treatment. Treatment numbers suggest that this offer was not made. The frequency of closed cast management varied substantially among sites (Table 4-1). Because the distributions of intervention by season, gender, and age are otherwise similar per site, however, it may be true that the enrollment instruction was observed for surgical treatments at each site during the time period of study participation. Also, the enrollment and rejection rates were similar among sites, providing additional evidence that each site attempted to enroll eligible patients.

Patients’ medical characteristics and behavior seem to be similar among sites. Comorbidities revealed an 8% incidence of patients receiving medical treatment for osteoporosis before fracture. This incidence was not significantly different among sites. All patients included in this report completed follow-up examinations. Interval 2 (6 weeks) participation was 72%, but participation decreased to 40% by 1 year (interval 5). The follow-up participation percentage was not statistically different among sites.

The combined average primary outcome measure (DASH score) was similar at all but one site. The dissimilar site seemed to restore function over a longer time interval. This finding was significant and could not be explained by average age or gender bias.

Primary outcome measure (DASH score) recovery was similar for A (extra-articular) fractures compared with C (intra-articular) fractures. Primary outcome measure (DASH score) recovery was faster for men compared with women. This trend was true at all intervals and approaches significance (Fig. 4-6).

FIGURE 4-6 Median DASH Scores. A, Median DASH score by fracture type at follow-up intervals 1-5. B, Median DASH score by gender at follow-up intervals 1-5.

Excluding bilateral injuries, grip strength recovery (compared with percentage of unaffected side), digital motion recovery (total active motion of affected hand fingers [measure the worst finger]), visual analog scale pain scores (0 to 10), and work status (coded in an ordinal manner as to ability to perform “original” work tasks) are the best clinical (physician-collected) values as correlated to DASH recovery (P < .001). Wrist motion does not correlate to DASH recovery. Forearm motion is important, but not to the same degree as grip strength recovery and digital motion preservation (Table 4-2).

Radiographic measurements were not generally predictive of DASH recovery. Only lateral tilt (volar or dorsal tilt captured in 5-degree increments on the true lateral view) moved significantly with DASH and only at interval 2.

Fixation differences by site and fracture type are listed in Table 4-3. Excluding patients with bilateral injuries, weak evidence suggests external fixation bridging is best early (interval 2 [6 weeks]) compared with volar fixed-angle plating (VFAP). At later (interval 3 [13 weeks] and beyond) follow-up, VFAP (no difference between manufacturers observed) is different (lower DASH) from external fixation bridging. As previously noted, cast care and primary “pin” fixation recover most slowly (Fig. 4-7).

FIGURE 4-7 DASH Score Recovery (Median). DASH score recovery (median) by interval (A) and sorted by treatment device (B).

Adverse events were recorded. Conversion from original intended treatment to a second type at a later date was considered an adverse event. Two scenarios were seen: (1) conversion from cast immobilization to internal fixation after loss of reduction (several cases) and (2) failure of original fixation to achieve reduction (one case). The adverse events in this series were sufficiently different and infrequent to prevent analyses.

Data were entered to the database from all sites via the Internet. In some instances, the data were entered contemporaneously with the episode of care (including operative documentation). Some of these data were directly incorporated into the electronic medical record present at the site of care using a VPN connection and HL-7 protocol. Most (70%) data were collected by investigators using patient-specific treatment, paper-based booklets, and entered into the database at some point after the visit.

One case is used to illustrate the data capture process in this study. A patient presented with a new closed fracture. The patient was an otherwise healthy English-speaking woman with no prior injury or chronic medical concerns. No nerve or tendon injuries were present. An unstable fracture with a dorsal-ulnar corner fragment and associated laxity of the distal radioulnar joint was diagnosed. Fixation maintained reduction of this intra-articular fracture and early active-assisted range of motion was initiated. DASH score recovery was progressive and nearly complete. The patient presented nearly 10 months after fixation with dorsal tenderness over a fixation bolt (adverse event) that resolved after removal of the screw only and no change in the DASH score (Figs. 4-8 to 4-10).

FIGURE 4-8 Three X-ray Views of an Unstable Distal Radius Fracture. A, Before treatment. B, After treatment using an IM nail.

FIGURE 4-9 Screen Shot of the “Patient Dashboard” Used to Manage Data Entry during Treatment and Follow-up.

FIGURE 4-10 Plot of Specific Patient’s DASH Score Recovery after Treatment.

Discussion

One goal of this study was to employ this Internet-based method of data collection to examine our first hypothesis:

As such, differences in primary outcome related to equipment factors are observed in this study, but are not large and seem to vary with time of analysis. Also, power analysis confirms that the number of patients required to detect meaningful and predictive difference between fixation types by manufacturer within specific groups (e.g., VFAP) cannot be estimated, with this data set, owing to the similarity of results by fixation manufacturer within fixation groups. Assuming that a difference might exist, our estimate suggests that each group would need to number in the thousands of patients to find any clinical difference between VFAPs. It would seem that a range of VFAPs are clinically identical.

External fixation (bridging) is different (trend) and better early during recovery. We believe this observation warrants special emphasis in the face of the number of plating products with strong messages being communicated to physicians via advertising and data reported to support VFAP published by Chung and associates.¹¹ Earlier work published by Kreder and coworkers¹² found no difference between fixation with plating versus external fixation for unstable fractures of the radius when the patients were randomly assigned to one of the two groups.¹² This article cannot be compared directly with the current work in that the plating techniques chosen for open fixation did not share the same characteristics as the techniques now presented to the market and represented within Chung’s work.

Proof that fixed-angle fixation of the radius would improve fixation rigidity using an in vitro extra-articular distal radius fracture fixation model was first shown by Gesensway and colleagues¹³ and later compared with fixation forces needed to allow early rehabilitation vis-à-vis a low load gripping (finger motion) model by Putnam and colleagues.¹⁴ Given generally understood knowledge of joint and ligament function, it is probably impossible to argue logically against restoration of function of all joints as early as possible. To be useful in the context of a utility/disutility analysis, however, at the outset of introducing alternative fixation methods into the matrix of fracture care, the instant and widely applicable benefits should be known. A specific analysis of distal radius fractures regarding utility/disutility has not been done. Inferring from the work of Gandjour and Lauterbach,¹⁵ one issue to be understood is the expected and observed volume of “greater benefit” for all distal radius patients who receive the best care compared with alternative (and, for argument’s sake, less expensive) care. In this example, “dollars available for health care” can be assumed to equal energy and are subject to the laws of thermodynamics, wherein they can neither be created nor be destroyed. To achieve utility, any increase in energy spent on radius fracture care should ideally result in greater societal health compared with spending these same dollars elsewhere.

Based on the lack of consistent differences in recovery speed seen in this study when comparing fixed-angle plates (any side or manufacturer), intramedullary rod fixation, and external fixation, and the lack of any formal comparative utility/cost analysis, we recommend that the surgeon should choose the fixation type to be employed based on factors other than device design.

A second goal of this study was to assess if:

Separating DASH averages by site was reassuring and surprising. As seen in Figure 4-11, the sites are more similar than different when the DASH alone is examined. One site does trend to recover more slowly, however. Ultimately the true test of such a display would be the ability to detect a best (or worst) process. Ideally, such a method could confirm a competent result (single event) and competent method (many events by many physicians). The concept of measuring and displaying surgical competence is gaining acceptance as a goal. The CUSUM (CUmulative SUM) method of presenting surgical success/competence data may prove reliable and acceptable to the users.^16,17

FIGURE 4-11 DASH Score Median. A, DASH score median with ranges at interval 2 showing site difference. B, DASH score median with ranges at interval 5 showing site difference.

Figure 4-12 shows examples of how physician-derived data and patient instrument might be combined to represent distal radius fracture recovery at specific intervals. To create this graph, the physician values that most closely correlated with DASH score (grip, finger range of motion, visual analog scale, and work status) were assigned values to create a physician index. This index score can be seen to move in the same direction and approximate proportion compared with the DASH instrument. The importance of this fact can be surmised from the difference in number of sites represented in Figures 4-11 and 4-12. Specifically, site 4 was not successful at collecting values for physician-derived data. Although site 4 did collect patient instrument data that could be analyzed to create Figure 4-11, the frequency and completeness of data collection were insufficient to enable inclusion of site 4 in Figure 4-12.

FIGURE 4-12 Plotting the Physician-Measures Index Score against the “Inverted” DASH Score. A-C, Plotting the physician-measures index score against the “inverted” DASH score for three sites at three intervals. The direct relationship between the DASH score and the physician-collected data is clear.

The importance of the DASH score accurately reflecting the movement/improvement in physician-derived data (physician-measures index) seems obvious. Assuming that stakeholders could agree on a desired standard percent of DASH score recovery without adverse event occurrence, a binary process could be constructed to represent expected recovery parameters. These data could enable the building of a CUSUM representation of care success without substantial change in the regular medical work process (i.e., limited physician data gathering requirement and patient data gathering via novel and inexpensive means [Internet]).

In this observational study, we intended to enroll “all comers” as related to distal radius fracture patients who met the inclusion/exclusion criteria. Strictly speaking, all patients presenting with a specific diagnosis should be enrolled. Enrollment numbers suggest, however, that patients receiving cast care as treatment were not enrolled as frequently as patients undergoing surgical intervention. Notwithstanding this limitation of the current study, for purposes of formulating some conclusions from the collected data set, the distribution of patient characteristics is encouraging in that it compares favorably with an earlier demographic study by Falch.¹⁸

Differences in primary outcome related to clinical factors were observed in this study. Some of these are significant. Taken together, it seems that using radiographs as a means of assessing fracture recovery is less meaningfully related to the primary outcome (DASH recovery) than comparative grip strength and finger motion. Forearm motion may be important in DASH score recovery, whereas pure wrist motion is not. That collecting the aforementioned data is less expensive and without x-radiation risk compared with routinely obtained radiographs is potentially important. If we assume that current practice is to obtain postsurgical radiographs at 2, 6, and 12 weeks; that each radiographic series costs $100 to the institution; that 300,000 distal radius events are treated in the United States per year; and that the 6-week and 12-week radiographs could be eliminated 90% of the time, the direct cost savings to the health care providers and institutions would be [(100 + 100) × 270,000 = US$54,000,000]. Although such a number is only three ten-thousandths of the U.S. health care budget, the value of identifying such opportunities is obvious.

Considering the need to conserve resources in medicine and the inability of orthopaedic equipment to possess sufficient embedded knowledge so as to make the physician’s decision making and physical skills of minimal importance, it seems obvious that true and ongoing tests of widespread efficiency (ability of each user to employ a tool effectively [desired outcome]) are needed with continuous comparison with an existing standard when new and differently priced tools are introduced to the market.¹⁹

Incomplete capture of late follow-up data occurred secondary to patients failing to return for later examination. The current data set has not been examined using an “intention to treat” method.²⁰ Presumably, this method would enable a better explanation of trends, although it may “downgrade” seemingly good results. In the absence of completing this analysis, the trend differences between sites and some equipment must be viewed with caution. The clinical correlations between DASH and grip and finger motion are so significant and largely dependent on earlier, more complete data that we expect these correlations to remain positive regardless of any new data points gained vis-à-vis our planned intention-to-treat analysis.

The database as used shows the Internet’s greatest strength: the ability to deploy and gather data with greatest efficiency and enforced accuracy. Widespread deployment would not be simple, however. Not everyone is ready and willing to understand individual effectiveness. Patients, who arguably have the most to gain, have almost no voice. Third parties (operating sites, payers, manufacturers) have variable and conflicting opinions ranging from “I am already doing it” (payers) to “I cannot do it” (operating sites). Physician leadership is stating that measurement is a necessity, but must deal with “push back” from many physicians who see quality measurement as an unfunded mandate or a “hidden agenda” or both.

In this study, we identified an ability to use a primary outcome instrument to observe individual and group recovery of function over time using a validated outcomes instrument (DASH) as the primary outcome measure. Separation of treated fractures into comparative groups is not likely to have occurred in a uniformly agreeable manner. Dahlen and colleagues²¹ describes the “under” classification of extra-articular fractures by plain radiographs. Because of this, these authors suggested classification following routine computed tomography scan after original radiographs. If operative treatment expense is an issue, however, imagine the utility, or disutility, of obtaining a computed tomography scan in all patients presenting with a distal radius fracture. That intraobserver and interobserver reliability of fracture classification is poor for any but the simplest classification scheme is also known.²² One clear opportunity, if the power of a clinical database distributed via the Internet is to have maximal benefit for clinical studies, occurs at the moment a disease or image is classified. Whether the physician is simply given enough information to become more consistent or, ideally, a machine algorithm is used to validate any crucial classification, the value to a clinical data collection and sorting process is evident.

Summary

As related to the primary outcome of distal radius fracture treatment, we have shown that:

This message moves against the differentiation sought by the medical equipment industry and the uniformity of payment for service enjoyed by insurers as it relates to professional service payments.

Pirzig² relates the story of a South Indian monkey whose hand is caught in a coconut with rice inside. The monkey cannot remove its hand when holding onto rice because his hand closed around the rice is larger than the hole he reached through with his open flat hand. There are hunters coming. So, the monkey must choose to forego his hunger or be caught. Maintaining his frame of reference, the monkey will lose his freedom, or worse.

Medicine’s situation is more problematic. The “hunters” may not be here tomorrow and so, depending on the variable ages and needs of physicians, the need to remove one’s “hand” from the “food” would be viewed differently. The trick becomes to see the need from the “eyes of the future.” From this vantage point, it might be possible to place the greatest value on the overall quality achieved and, in this context, the need for patient-validated measurement of equipment, physicians, and sites is apparent. The database method presented fills the need to compare variables in a manner that enables an ongoing understanding of real-world efficiency and simultaneously provides individual understanding of effectiveness.

ACKNOWLEDGMENTS

The authors acknowledge N. J. Meyer, MD, R. Strauch, MD, and C. Jennings, MD, for working to refine the data set and allowing their cases to be included, and J. Agel and P. Lender for working to analyze the data set and providing monitoring and maintenance of the institutional review board status.

Dr. Putnam has an ownership interest in the Database method (BoundaryMedical) employed by this study with oversight supplied by the University of Minnesota. Dr. Putnam has a royalty interest in distal radius fracture fixation products sold by SBI and Wright Medical Technology.