Building Quality Improvement Initiatives through Anti-Cancer Stewardship in the Veterans Health Administration (VHA): CML Pilot

The Department of Veterans Affairs (VA) Pharmacy Benefits Management (PBM) Services created a national stewardship with focus on anti-cancer therapeutics within a disease-specific cohort . Directing stewardship initiatives to a cohort of Veterans with a common malignancy is a novel approach.

A multidisciplinary team with National PBM and VISNs (Veterans Integrated Service Network) promotes standardization of practice with accountability to monitor outcomes as a continual quality improvement process. A toolkit was built to support stewardship activities and includes:

• near-real-time process to identify a national disease-specific Veteran cohort

• disease validation process

• centralized MUE data collection

• educational resources

• site for communication and training materials

Tyrosine Kinase Inhibitors (TKIs) account for the second highest utilization among outpatient oral anti-cancer therapies in VHA . Imatinib has demonstrated persistent efficacy and is the preferred TKI on the VA National Formulary.

Utilizing our toolkit, we developed two MUEs for our national CML cohort to determine why Veterans were not prescribed imatinib as 1L therapy and reasons for switching from first to second line therapy.

Methods

Business rules identified VISN CML cohorts by using data sources and diagnostic coding from the VA Electronic Health Record (EHR): primary care and hematology/oncology visits (ICD-10), inpatient stays (ICD-10), and problem lists (ICD-10; SNOMED CT). Retrospective chart review of two subsets:

• Veterans who received a drug other than imatinib as their initial VA prescription

• Veterans who received at least two unique drugs to treat CML

Data abstracted for MUE #1:

• Disposition of TKI prescription

• If new initiation, documented reason(s) for drug selection

• If continuation, prior drug therapy history

Data abstracted for MUE #2:

• Documented reason(s) for switch to 2L

• If adverse drug reaction (ADR) included as a reason, is it documented in the EHR?

Analysis combined responses with prescription data from the EHR.

Results

Of the 18 VISNs, 9 participated in MUE data collection. 365 patients were identified as not having received imatinib 1L and 360 were identified as having received at least two unique drugs to treat CML. These counts are each approximately 30% of the total CML patient cohort. In the total cohort, ~50% of patients received imatinib 1L and have not switched to 2L.

Responses for 1L TKI therapy (N=365)

• 199 (55%) continued therapy originally started outside VA with 67% of these patients (n=133) without documentation of prior imatinib therapy

• 166 (45%) were newly initiating CML therapy with top documented reasons for drug selection:

  • Intermediate or high risk (e.g., SOKAL risk scoring): 77 (46%)

  • 46 (28%)

  • 18 (11%)

  • Dasatinib was the most prescribed 1L therapy in this cohort (64%; 235), followed by nilotinib (23%; 85), which had a higher prevalence when therapy was continuation (30%).

  • 53 patients (15%) had subsequent imatinib therapy

  • Responses for switch from 1L to 2L therapy (N=360)

    • Top documented reasons for switch from 1L to 2L therapy

      • 215 (60%)

      • 81 (23%)

      • Resistance / inadequate response: 69 (19%)

    • Only 37 (17%) who experienced an ADR had it documented in the EHR

    • The most common 2L drug was dasatinib (55%;198). The median number of days between 1L and 2L initiation was 275 (range: 7 - 2, 573)

    • 136 patients (38%) received third line therapy or beyond

Conclusions

A formalized oncology stewardship with focus on disease-specific initiatives within a national healthcare system is feasible. Responses from our 1L MUE indicate that despite geography, imatinib remains the preferred TKI and accounts for the highest utilization. Most veterans not receiving imatinib as 1L therapy, received dasatinib or nilotinib. Continuation of therapy due to transition of care to VA was the primary reason for 1L therapy with a non-preferred TKI. Other reasons included: FDA indication and SOKAL risk scoring.

In our pilot 1L to 2L MUE, ADRs were the most common reason for CML patients to be switched to 2L (45%) followed by resistance/inadequate response and disease progression . Most ADRs were not documented in the EHR.

Further interventions in TKI prescribing may be instituted in select sites . Following initiatives for CML, Anti-Cancer Stewardship will be expanded to other oncologic diseases.