Physician Knowledge, Practice Patterns, and Barriers Encountered Regarding Guideline-Concordant Use of Bone Modifying Agents for Prostate Cancer

Abstract

Background:

Guidelines recommend bone-modifying agents (BMAs) for patients with castrate-resistant prostate cancer (CRPC) and bone metastasis, but not for castrate-sensitive prostate cancer (CSPC). Physicians beliefs and practices regarding BMA therapy are poorly understood.

Methods:

This was a qualitative interview study with embedded Likert-scale elements. Study participants were physicians who treat prostate cancer, located within an academic cancer center or an affiliated community-based network. Participants were asked about their experiences and practice patterns regarding BMA therapy. Participants used Likert-scale items to identify the most common barriers to guideline-concordant BMA use and the most effective potential interventions. Participants were subsequently asked to rank the 3 most common barriers and the 3 most effective interventions to reduce underuse (for CRPC) and overuse (for CSPC).

Results:

19 physicians were invited and 15 participated; 1 physician did not answer some questions as outside of their practice scope. All were aware of the recommendation for BMAs in CRPC. 14% (2/14) were unaware of the recommendation against BMA use for CSPC; an additional 29% (4/14) believed that BMA use could be appropriate for CSPC depending on the metastatic disease burden. 36% (5/14) were unaware of recommendations for screening and treatment of low bone mineral density. The most common barriers (occurring “often” or “sometimes”) were obtaining dental clearance (11/15) and insufficient clinic time (6/15). The interventions identified as most effective to reduce underuse were dental navigation (11/15) and EMR-based guidance (9/15). The interventions identified as most effective to reduce overuse were peer-to-peer education (14/15) and EMR-based guidance (13/15).

Conclusions:

Awareness of guideline recommendations for screening and treatment of low bone mineral density and against BMA use for CSPC was good, but not complete. Dental navigation, peer-to-peer education, and EMR-based guidance were preferred intervention strategies to improve guideline-concordant use.

Introduction

Prostate cancer causes over 34,000 deaths annually, placing it among the deadliest cancers for US men. Bone metastases frequently contribute to prostate cancer morbidity by causing skeletal-related events (SREs): pathologic fracture, spinal cord compression, and bone pain requiring intervention. Among metastatic prostate cancer patients, 30–50% will experience SREs,^1,2 with substantial loss of function and quality of life.³

National Comprehensive Cancer Network (NCCN) Guidelines recommend the bone-modifying agents (BMAs) zoledronic acid (Zometa) and denosumab (Xgeva) to prevent SREs among patients with metastatic, castration-resistant prostate cancer (mCRPC) and bone involvement.^4–6 BMAs are also recommended to prevent bone mineral density loss and osteoporotic fractures among patients receiving androgen deprivation therapy; NCCN Guidelines recommend screening for and treatment of osteoporosis or osteopenia with high fracture risk.^6,7 However, randomized clinical trials have found no benefit from early initiation of zoledronic acid among metastatic, castration-sensitive prostate cancer (mCSPC) patients,^8,9 and both NCCN and ESMO Guidelines therefore do not recommend BMAs therapy in mCSPC.^6,10

Studies have documented the dual problems of BMA underuse and overuse. Approximately one third of Medicare patients with mCRPC do not receive BMAs (underuse),¹¹ while one quarter of those with mCSPC and no evidence of osteoporosis/osteopenia receive BMAs unnecessarily (overuse).¹² Little is known about the clinical decision-making factors that contribute to these gaps in guideline-concordant care. A Swiss survey of physicians identified perceived osteonecrosis risk and insufficient clinic time as the most frequently-cited reasons for underuse,¹³ but this has not been evaluated in the US setting. Our study’s goal was to describe oncologists’ beliefs, practice patterns, and perceived barriers to guideline-concordant use of BMAs for prostate cancer.

Methods

Physicians (MD, DO, MBBS, or equivalent) treating prostate cancer at an academic cancer center or two affiliated community-based oncology networks were eligible to participate. Physicians were recruited via email (Supplementary Material 1) through consecutive sampling. An experienced moderator (SC) led 30- to 60-minute interviews following a semi-structured guide (Supplementary Material 2) from July through October 2022 until thematic saturation was achieved. Interviews were conducted and recorded remotely via Zoom. Participants received $100 gift cards. Automated transcription via Microsoft Office was followed by 100% manual review/correction. Recordings were deleted after transcription.

Interviews began with participants describing their own practice regarding BMAs, including patients considered for treatment and information sources guiding practice. Participants were asked which BMA they generally prefer and which drug characteristics (e.g. convenience, cost) factored into their preference. Participants then considered 3 NCCN recommendations related to BMAs: 1) use for SRE prevention in CRPC, 2) non-use for SRE prevention in CSPC, 3) bone mineral density evaluation and treatment. For each recommendation, they related whether they had prior knowledge of it, how often and why they departed from it, and whether “grey areas” exist where it may not apply.

Next, participants viewed a list of 7 potential barriers to BMA therapy, selecting Likert-scale options for how frequently each occurred: “often,” “sometimes,” “rarely,” or “never.” Participants were asked to identify additional barriers and choose the 3 most and 3 least common.

Finally, participants considered 8 potential interventions (Supplementary Table 1) for addressing underuse, assessing each as “very likely,” “somewhat likely,” “somewhat unlikely,” or “very unlikely” to increase appropriate BMA use in patients with mCRPC and selecting the 3 most likely and 3 least likely to be effective. Finally, they considered the 3 most likely to be effective in addressing overuse (reducing inappropriate BMA utilization for patients with CSPC.)

Categorical responses (Likert-scale choices and most/least ratings) were analyzed in Excel. Transcripts were analyzed using QDA Miner software. Three authors (AM, SP, SC) developed an initial coding scheme. One transcript was independently coded by 2 authors (SP and SC), with discussions to refine the scheme and resolve discrepancies. This process was repeated with a second transcript, establishing a high level of intercoder reliability (97.6%, Kappa=0.952). The remaining transcripts were coded by one of two authors (SP and SC). This study was approved by the MSK IRB.

Results

I. Participant characteristics.

Of 19 physicians invited, 15 participated: twelve (80%) medical oncologists and 3 (20%) urologists. Most (13, 87%) were men. The mean years in practice was 15.4 (range 3–48). Five were located at the academic cancer center, and 10 practiced within one of the community-based networks (Table 1).

Table 1.

Characteristics of interview participants.

Characteristics	N = 15
Mean number of years in practice (range)	15.4 (3–48)
N (%) Female	2 (13)
N (%) Male	13 (87)
Specialty, N (%)
Medical oncology	12 (80)
Urology	3 (20)

II. Key considerations when using BMAs

When asked about their practice patterns around BMA use, many participants emphasized guideline recommendations. They commonly discussed CRPC with bone metastases and low bone mineral density while receiving androgen deprivation therapy as settings in which they recommend BMA therapy.

Regarding information sources that participants relied upon, NCCN Guidelines were by far the most commonly identified. Others included the American Society of Clinical Oncology (ASCO), UpToDate, TheMedNet, Research to Practice, Vumedi, and the scientific literature.

Patient medical history was another key factor in therapeutic decisions. Baseline renal function and dentition were the most commonly discussed factors, with concern about precipitating osteonecrosis of the jaw in patients with poor dentition:

“The major information that I’m concerned about is…the dental aspect, because that is what’s going to lead, if unrecognized, to osteonecrosis of the jaw.”

III. Key concepts in selecting BMAs

Most participants reported using both denosumab and zoledronic acid depending on the clinical circumstances. However, most expressed a general preference for denosumab; only 3 (20%) reported using zoledronic acid more frequently. These 3 identified several advantages to this drug, including less frequent dosing:

“For most patients it’s more convenient to have an every-three-month infusion as opposed to a monthly injection. I think that calculus is pretty clear.”

Lower incidence of hypocalcemia was another benefit with zoledronic acid:

“[With] a heavy bone met burden, I’m more worried about hypocalcemia [and] tend to favor Zometa.”

Additional advantages were lower incidence of “rebound” fracture and lower price.

Among those favoring denosumab, convenience was the primary reason. Here, many participants pointed to the route of administration and less-intensive monitoring for kidney injury:

“It’s easier for the patient to get a subcutaneous injection as compared to sitting in a chair and get infusion for 15 minutes. And then the kidney part is always in mind.”

Notably, however, some participants saw the 2 drugs as equally convenient because they had adopted a 12-week dosing interval for denosumab:

“I think we do have enough data…that quarterly denosumab for prostate cancer patients is pretty non-inferior to the monthly dosing…So I use quarterly denosumab.”

Financial considerations played little part in participants’ preferences. None described greater reimbursement for denosumab as an advantage, and many reported being unaware of this issue:

“I’ve never even really researched [it]. So I don’t even know which would provide greater reimbursement.”

Similarly, most never considered patients’ out-of-pocket costs:

“I have not heard anybody have a high out-of-pocket cost for these medications.”

“That, to my knowledge, has not been an issue for my patients.”

The sole financial factor influencing the choice of BMAs was insurance coverage, typically necessitating a change to zoledronic acid:

“…if the insurance company ever tells me that I have to use [zoledronic acid] over [denosumab] for treating bone metastases, I’m not going to question it.”

Few participants were aware that denosumab is slightly more effective in delaying SREs compared to zoledronic acid, so this was not a factor in drug selection - even among those preferring denosumab:

“I’m actually not familiar with that literature…my understanding…is that there’s no dramatic difference between the two…”

“No, I’m not aware of this data, but I have been using denosumab.”

IV. NCCN Recommendation #1

All participants were aware of NCCN Guidelines recommending BMAs for patients with CRPC and bone metastases. Estimates of departure rates ranged from “never” to 30% of cases. However, this difference may be explained by some participants considering patient medical contraindications or non-adherence as departures from guidelines leading to higher self-reported departure, while others felt that they had adhered to the guideline recommendations in such cases leading to lower self-reported departure.

Several participants identified low metastatic burden and/or perceived low fracture risk as an appropriate reason to deviate from this recommendation:

“For a patient with low volume whose disease is responding well…their risk of fracture is probably not substantially changed by the addition of an agent like this.”

Limited life expectancy was another “grey area”:

“…it might [not] apply to somebody who’s very elderly and has a short life expectancy for other reasons”

V. NCCN Recommendation #2

Most participants (12 of 14) were aware that BMAs are not recommended in the CSPC setting. Two were unaware, and one additional participant was aware of the recommendation but believed that it applied only to zoledronic acid and that denosumab was recommended for CSPC.

Among those aware of this recommendation, 4 believed that a high burden of bone metastasis and/or clinician judgement of high fracture risk were appropriate reasons to depart from it:

“For someone who has [CSPC], I do not use bisphosphonates. Unless they have extensive bony metastasis and I’m very worried about their bone health.”

VI. NCCN Recommendation #3

Most participants (11 of 14) were aware of the NCCN recommendation for baseline bone mineral density testing and BMA treatment for patients with osteoporosis or osteopenia and high estimated fracture risk. Self-reported deviation from this recommendation varied more widely, from “never” to “probably a lot.”

Among those having deviated from this recommendation, time pressure and prioritization of other clinical needs were common themes:

“Honestly it’s just the sheer amount that needs to happen at the initial visit and potentially not addressing it later on. It’s just sheer burden of moving parts.”

Other participants expressed a lower level of attention to this recommendation:

“[I deviate] probably a lot…I need to do better as far as…doing more of the testing and for osteoporosis, osteopenia. And be on it a little bit more. That’s definitely a deficiency.”

When asked about “grey areas,” participants commonly saw the time frame for this recommendation as flexible, wherein bone mineral density can be addressed non-urgently:

“I do get baseline DEXAs on my patients that are starting on hormones, but honestly…it’s not a priority for me when I’m first staging them for their cancer or first getting them onto therapy. I try to do it within the first year.”

Others believed that bone density screening might be unnecessary for patients with anticipated short duration of ADT, due to perceived lower risk for osteoporosis:

“I generally will not do this for patients getting four or six months of ADT.”

VII. Barriers to BMA therapy

The most commonly encountered barrier to appropriate BMA use was patient dental disease (12/14 reported “sometimes” or “often”), with nearly all participants (14/15) ranking this as a top-3 barrier (Figure 1, Supplemental Table 2, Supplemental Table 3).

“Definitely a barrier because some patients I can’t even get to go to the dentist and I’m really reluctant to put this medicine on someone who has very poor teeth.”

Figure 1:

Participant assessments of potential barriers to appropriate use of BMA therapy among patients with castration resistant prostate cancer (N = 15). Participants were asked how often they have encountered this barrier in their clinical practice? *Not all participants provided responses for each potential barrier, resulting in fewer than 15 responses in some cases.

The second most common barrier was insufficient time in clinic to discuss risks and benefits of BMAs (6/14 “sometimes” or “often”):

“Most doctors get paid by being a widget maker and how much they do…in the least time…We are stretched thin…it’s a barrier.”

The least commonly encountered barrier was logistical barriers in clinic (0/14 “sometimes” or “often”).

VI. Interventions to improve utilization

Some proposed interventions were broadly popular among participants (Figure 2, Supplemental Tables 4 and 5), including dental care navigation (14/15 “somewhat” or “very” likely to increase BMA use):

“That would help a lot…Our dental care system is terrible…patients have a really hard time…it’s pretty complicated.”

Figure 2:

Participant rankings of potential strategies to increase BMA utilization among patients with metastatic castration resistant prostate cancer. Participants were asked to judge the three most effective and the three least effective among the proposed strategies. Among the three strategies identified as most effective, we scored them as 1^st, 2^nd, and 3^rd choices if the participant stated such an order, or if the participant gave no order then all were scored as 2^nd choice. Similarly, among the three strategies identified as least effective, we scored as 5^th, 6^th, and 7^th choices if an order was given, or as 6^th choice if not. Strategies which were not mentioned as being among either the three most or least effective were scored as “neutral.”

Twelve participants rated personalized feedback as somewhat or very likely to be effective:

“When you show people what you’re supposed to be doing and everybody else is doing it and you’re not doing it, that’s powerful.”

Patient-specific notifications were similarly popular (12/15 “somewhat” or “very” likely):

“If somebody said to me, these are the people who are candidates for it, who you’re not treating, I would appreciate that.”

Participants diverged around other proposed interventions. For example, while 6 saw patient financial navigation as somewhat or very likely to work, 9 saw it as somewhat or very unlikely, often because they doubted patients faced financial barriers:

“Because I haven’t seen any financial barriers, I would say very unlikely.”

Regarding overuse of BMAs among patients with castration-sensitive disease, participants almost uniformly favored peer education and EMR-based guidance (14/15 and 13/15, respectively, ranked among top 3) (Figure 3, Supplemental Table 6):

“[For] overutilization…peer physician education…would be most helpful.”

“If I’m [starting a BMA for] someone who’s castrate sensitive and I got a pop-up, I would probably be like, ‘wait – that’s right!’”

Figure 3:

Participant rankings of potential strategies to decrease BMA utilization among patients with metastatic castration sensitive prostate cancer. Participants were asked to judge the three most effective among the proposed strategies. Among the three strategies identified as most effective, we scored them as 1^st, 2^nd, and 3^rd choices if the participant stated such an order, or if the participant gave no order then all were scored as 2^nd choice. **Dental care navigation, financial navigation, and patient education were not ranked by any participant, and therefore were omitted for clarity.

Discussion

To our knowledge, this is the first in-depth study of US-based physicians’ attitudes and practices regarding BMA therapy for prostate cancer. Despite good awareness of practice guidelines, participants reported frequently encountering barriers to appropriate BMA utilization.

All participants were aware of NCCN recommendations for the use of BMAs in CRPC. However, adherence was limited by patient and practice factors, most notably dental disease (and the resultant need for pre-treatment dental care and clearance) and time pressure in clinic. Dental care presented a barrier to treatment despite the availability of integrated, within-network dentists, and dental “navigation” – assistance to patients in scheduling and keeping dental care appointments – was identified consistently as the intervention most likely to improve appropriate BMA underuse.

Notably, the barriers reported by our participants overlap with those within the Swiss healthcare setting.¹³ When surveyed about reasons for BMA non-initiation in cases where initiation was recommended, Swiss physicians most commonly identified time constraints (35%), prioritization of treating the primary tumor (thematically similar to our study’s barrier of “prioritization of other clinical needs”) (33%), and risk of ONJ (20%).

Not all participants were aware that BMAs are not recommended for patients with CSPC. Additionally, several who reported being aware believed that early initiation of BMA therapy may be appropriate in cases of CSPC with high burden of bone metastasis, though guidelines make no recommendation for patient selection on this basis.

In contrast to the system-level barriers and interventions identified as most effective in increasing BMA use in CRPC, participants generally felt that educational interventions would be the most effective in reducing BMA overuse in CSPC. This is in line with their recognition that such use may represent a knowledge deficit.

Unexpectedly, we found an absence of financial considerations in choosing between these agents. Participants reported being unaware of differences in practice reimbursement, and that this played no role in their drug selection. Also unexpected was the absence of patient-level “financial toxicity” as a barrier, especially given denosumab’s high price: approximately $2,700 per dose vs. $33 for zoledronic acid.¹⁴ While this likely reflects the fact that insurance coverage shields most patients from most of these costs, it is also possible that patients who do face substantial out-of-pocket costs rarely bring these concerns to their oncologists’ attention.¹⁵ Another unexpected finding was that no participants identified head-to-head trial data as favoring denosumab.⁴

In prior work using discrete-choice experiments, physicians identified effectiveness in SRE prevention and minimizing patient out-of-pocket burden as the most advantageous qualities of BMA drugs.¹⁶ However, while our participants still viewed these as advantageous qualities, these qualities did not influence our participants’ drug selection because they viewed denosumab and zoledronic acid to be effectively equivalent in SRE prevention and patient cost burden. In contrast, physicians in the discreet choice experiment identified a convenient route of administration as a relatively minor advantage,¹⁶ but this factor appeared to sway many of our participants towards denosumab in absence of other, more substantial, perceived differences. Interestingly, in a similar discreet choice experiment, cancer patients had no preference between subcutaneous and intravenous administration, suggesting that increased real-world use of zoledronic acid would be acceptable to patients.¹⁷

Some interventions favored by our participants may be time and resource intensive. Dental care navigation, the preferred intervention for BMA underuse, would likely require additional staffing or administrative burden on existing staff. However, several participants favored EMR-based guidance, a strategy recently employed in a Veterans-Affairs-based setting with substantial improvement in BMA utilization.¹⁸ Thus, there may be leeway to improve BMA utilization through comparatively low-cost changes such as workflow redesign.¹⁹

Regarding BMA therapy in CSPC, our understanding of effective strategies to prevent medication overuse in oncology is in its early stages, and relatively little data exist.^20,21 However, our participants were open to interventions including peer education, audit and feedback, and EMR-based decision-support tools, all of which are evidence-based in other healthcare settings.²²

Though our study included physicians in both academic and community settings, generalizability presents a limitation. The community oncologists were part of regional networks, and physicians’ beliefs and behaviors may differ in solo or unaffiliated practices. For example, reimbursement may be an important consideration in physician-owned practices, and dental barriers may be even greater for practices without integrated dental care.