The Dobson DaVanzo Report: What It Says and Why It Matters

Home > Articles > The Dobson DaVanzo Report: What It Says and Why It Matters
By Phil Stevens, MEd, CPO, FAAOP

In today's healthcare climate, cost matters. As healthcare expenses continue to rise, third-party payers increasingly scrutinize the cost-to-value ratio of covered services. As O&P clinicians, we see and understand the functional value of the services we provide, including improvements in the areas of gait, balance, and the ability to accomplish activities of daily living. However, there is an additional economic value to these same services. In our era of preventive medicine, insurance companies understand that some healthcare services can be seen as investments, ultimately decreasing an individual's healthcare costs over an extended period of time. A report, written by Dobson DaVanzo & Associates, Vienna, Virginia, commissioned by the Amputee Coalition with funding from the American Orthotic & Prosthetic Association (AOPA), provides a retrospective analysis of Medicare claims data to determine if subjects who received certain O&P services ultimately experienced reduced healthcare utilization compared to closely matched, untreated subjects.

In short, the analysis found that even with the added costs of the interventions themselves, within a short time overall healthcare costs were ultimately decreased among those who had received lower-limb orthoses, and the healthcare costs were comparable among those who had received spinal orthoses or lower-limb prostheses. This suggests that the initial costs of the O&P services largely paid for themselves in the ensuing months by reducing healthcare utilization in other areas. Clinicians should be aware of the idea that O&P services can be seen as a healthcare investment. This article reviews the details of the Dobson DaVanzo report and provides a glimpse into the impact O&P services have on subsequent healthcare costs, such as fractures and falls, emergency room (ER) admissions, and acute care hospitalizations.


The effort began when AOPA and the Amputee Coalition purchased a custom data set from the Centers for Medicare & Medicaid Services (CMS). Drawing from healthcare claims submitted during the 18 months between January 2008 and June 2009, the resulting study group consisted of patients who received certain O&P services. An exhaustive review of the included L-Codes is beyond the scope of this article; however, for the subgroup of patients who received lower-limb orthoses, the list of interventions included custom AFOs (L-1940, L-1960, and L-1970); noncustom AFOs (L-1930 and L-1932); custom KAFOs (L-2036); and both custom and noncustom knee orthoses (L-1843, L-1844, L-1845, and L-1846).

The subjects also had to present with one of several defined etiological diagnoses. These included, but were not limited to, nontraumatic joint disorders, osteoarthritis, acute cerebrovascular disease, late effects of cerebrovascular disease, acquired foot deformities, paralysis, multiple sclerosis, and spinal cord injury.

Once these beneficiaries were identified, CMS provided information on all of their healthcare claims between 2007 and 2010, yielding a year of claims prior to receipt of their lowerlimb orthoses and at least 18 months subsequent to receipt of their devices. These claims were drawn from all major care settings including inpatient and outpatient hospital visits; longterm care hospitalization; skilled nursing facilities (SNFs); inpatient rehabilitation facilities; home health agencies; hospices; physician visits; and durable medical equipment, prosthetics, orthotics, and supplies (DMEPOS) providers.

In addition to defining the study group and their healthcare utilization rates, a comparison group of patients who did not receive O&P services was also generated. These individuals were drawn from healthcare claims across the same care settings and within the same time period as the study group. Prospective comparison group subjects were matched to the subjects in the study group based on diagnosis, gender, age, and state of residence. The initial data set from CMS identified five potential comparison subjects for each patient in the experimental group.

The authors of the study then used propensity score matching techniques to match individual comparison subjects to individual study subjects. These techniques are well defined in academic literature and are used when randomized controlled trials (RCTs) are not possible or are impractical to administer. Their aim is to remove as much selection bias as possible in assembling study and comparison groups during a retrospective analysis, yielding observations that attempt to mimic RCTs. Going beyond the initial matching efforts of CMS, the propensity scores were used to more rigorously identify matching subjects in each group based on diagnosis, comorbidities, age, gender, race, and historical healthcare utilization in the year prior to the etiological diagnosis that precipitated receiving O&P services.


Using these techniques, the data from CMS yielded 34,864 carefully matched, highly similar subject pairs with one representing the study group and one representing the comparison group. In considering the 18-month window subsequent to receipt of their lower-limb orthoses, the average Medicare payment for the study subjects was $27,007, or an average of $1,500 per member, per month (PMPM). During the same time period, the average Medicare payment for the comparison subjects was $29,927, or an average of $1,663 PMPM-$163 more per comparison subject, per month. Thus, healthcare costs for those subjects who were treated with a lower-limb orthosis were about 10 percent lower than for those in the comparison group.

These figures take into account the reality that patients who receive such devices require some level of training with them, so it was not surprising to see that the average number of days in an inpatient rehabilitation facility was higher for the study group (0.72) than for the comparison group (0.52). An even more dramatic difference was seen in the use of outpatient therapy, where the study group averaged 17.4 visits compared to an average 12.1 visits by the comparison group. Given the costs associated with the provision of the lowerlimb orthoses and the additional costs of inpatient and outpatient rehabilitation, all of which would seem to suggest higher healthcare utilization rates among the study group, it is reasonable to ask from where the savings came. Importantly, the authors of the study examined this area as well.

The CMS data revealed that the comparison group had higher healthcare utilization rates due to fractures and falls, ER admissions, and acute care hospitalizations. Where the average fractures and falls among study subjects was 1.45 events, there was an insignificant increase observed in the comparison group with 1.52 events. There was a significant increase in the average number of ER admissions experienced by the comparison group (1.20) compared to that observed in the study group (1.08). A similar difference was observed with respect to acute care hospitalizations, where the utilization of the comparison group (0.70) exceeded that of the study group (0.62).

The PMPM expense provides additional insights about the differences between the two groups. There were areas where the healthcare costs of the study group exceeded those of the comparison group. Most noticeably, the average PMPM expense for durable medical equipment (DME) among the study group exceeded that of the comparison group by $58. Similarly, the PMPM expense for inpatient rehabilitation among the study group averaged $14 more than for that of the comparison group. However, these need to be considered against those areas where the average PMPM expenses were higher in the comparison group. These included acute care hospitalizations ($128), SNF expenses ($43), and long-term care costs ($15). These numbers provide valuable clinical insights into the mechanisms that help explain the net healthcare cost savings that appear to occur when a lower-limb orthosis is received.


As with lower-limb orthoses, a study group was identified according to predetermined Current Procedural Terminology (CPT) codes reflecting a lower-limb amputation and the provision of any of several L-Codes for immediate postsurgical, preparatory, and definitive prostheses for Symes, transtibial, and transfemoral amputees. Consistent with the stated goals of the investigations, subjects were only selected if their amputation occurred within the 12 months previous to receipt of their prostheses. Using the same risk-adjusting matching techniques described earlier, a comparison group was similarly assembled. Those subjects for whom an adequate comparison match could not be found were excluded from further consideration. Ultimately, 428 highly matched pairs were selected to represent the study and comparison groups respectively.


Given the higher costs associated with lower-limb prostheses, the subsequent savings in healthcare expenses experienced by those in the study group were not able to fully offset the costs of the prostheses. However, the total costs for the two groups were remarkably close. At the conclusion of the 12 months that followed the receipt of their prostheses, the healthcare utilization costs within the study group were within 1 percent of those of the comparison group over a comparable time span. That is, the costs of the initial prosthesis were, on average, nearly amortized within a year of its provision.

As with the lower-limb orthosis cohorts, additional insights can be drawn from the PMPM figures in the lower-limb prosthetic group. For the study group, the largest medical expense was predictably attributable to the prosthesis. For these individuals, the average PMPM DMEPOS cost was $1,554, or roughly seven times more than that of the comparison group. A smaller addition in healthcare utilization costs was encountered for outpatient physical therapy where the PMPM cost of the study group averaged about $173 more than that of the comparison group (utilizing, on average, about two more outpatient therapy visits).

However, these differentials were largely offset by the increase in acute care hospitalizations experienced within the comparison group. In this group, the average incidence of such hospitalization was reported at 1.51 visits, compared to an average of 1.18 visits in the study group. Translated into healthcare dollars, this equated to a PMPM increase in healthcare costs of $839 for those individuals in the comparison group. Similarly, there was a moderate increase in physician visits by those individuals in the comparison group, which created an additional increased PMPM cost of $341.

Further cost differentials were encountered as individuals in the study group appeared to be more likely to live at home and receive outpatient therapy, while their counterparts in the comparison group were more likely to receive facility-based care, such as inpatient rehabilitation. The increased utilization of these and other inpatient services created further increased PMPM costs of $256 for the individuals in the comparison group. Similarly, while the utilization of SNFs were described by the authors as statistically comparable between the two groups, there was a slightly higher utilization rate by those individuals in the comparison group, with an additional modest increase in associated healthcare costs.

When considering some of the negative events that could be tracked using the CMS data, the results were mixed. The average incidents of fractures and falls in the study group (0.90) was slightly higher than that observed in the comparison group (0.72), presumably as a result of the increased risks that accompany increased physical activity. However, the average incidence of ER admissions for the comparison group (2.10) was significantly higher than that experienced by the study group (1.55).


The CMS data also allowed the study authors to investigate the apparent impact of K-levels on overall healthcare utilization rates. Perhaps surprisingly, when prosthetic users were broken down into low-level device users (Medicare Functional Ambulation Levels K1 and K2) and high-level device users (Medicare Functional Ambulation Levels K3 and K4), their average 12-month healthcare utilization costs were actually comparable, with those of K1/K2 individuals being slightly higher. Once again, the increases in initial prosthetic costs between the two groups were offset by the subsequent differences in other healthcare utilization rates.

The largest contributor to this offset was found in the utilization of SNFs, where the rate for K1/K2 level individuals (2.02 average admissions) exceeded that for the K3/K4 level individuals (0.84), creating a PMPM cost disparity of $401 that largely offset the costs associated with the more expensive prostheses. Additional costs associated with home health services and acute care hospitalizations among the K1/K2 subgroup further accounted for differences in overall healthcare utilization rates. Fracture and fall rates and ER admissions for the two subgroups were comparable. Though different mechanisms of injury between the two groups might be inferred, these cannot be determined from the published data.


Clinicians need to be aware of the findings of the Dobson DaVanzo report. In a healthcare system that scrutinizes every expense, it is important for practitioners to be aware of the decreased overall healthcare utilization rates that occur when O&P services are provided. In the case of lower-limb orthoses, these net savings are immediate, with fewer acute care hospitalizations and ER admissions leading to reduced Medicare-funded healthcare utilization. In the case of lower-limb prostheses, the increased initial costs require about a year before they are essentially amortized by the savings that accompany decreases in the use of facility-based care centers, acute care hospitalizations, and ER admissions. Importantly, these immediate and eventual cost savings occur in addition to the improved functional abilities that accompany the receipt of an appropriate lower-limb orthosis or prosthesis. Such talking points, backed by a rigorous analysis of CMS-provided data, allow clinicians to defend the value of our interventions in an increasingly cost-conscious and skeptical healthcare environment.

Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City, Utah. He can be reached at

Editor's note: To read the complete report, visit