Assessment of Financial Difficulty in Participants with Chronic Lymphocytic Leukemia and Multiple Myeloma

Financial difficulties will be assessed using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30) financial difficulty item #28 regarding financial difficulties with a modified recall period: "Has your physical condition or medical treatment caused you financial difficulties in the past year?", measured as "Not at all/A little/Quite a bit/Very much." Participant responses will be dichotomized as follows: "Not at all" or "A little" classified as "No", and "Quite a bit" and "very much" classified as "Yes." This dichotomized response will be assessed using Wilson score confidence interval (95%).

The association of patient report of financial difficulty as measured by the EORTC QLQ-C30 item #28 regarding financial difficulties with a modified recall period: "Has your physical condition or medical treatment caused you financial difficulties in the past year?" with insurance status (defined as those with Medicare/Medicaid v. those with commercial insurance) will be assessed using Mann-Whitney U test. Item #28 on the EORTC QLQ-C30 is measured as "Not at all/A little/Quite a bit/Very much", converted to 0, 1, 2 and 3 respectively for the purposes of analysis.

The association of patient report of financial difficulty as measured by the EORTC QLQ-C30 item #28 with receiving treatment at sites of care that report offering patients financial guidance through navigators or social workers as measured by question 20 of the protocol practice survey will be tested using multinomial logistic regression modeling. In this model the outcome will be the patient report of financial difficulty, and the covariates will be the financial guidance questions from the protocol practice survey, along with disease/treatment characteristics and indicators of patient socioeconomic status. Question 20 of the protocol practice survey asks the site to state the financial navigation services that are currently offered at their practice to all cancer patients or their families. If a site offers at least one service, they will be treated as offering financial guidance (measured as "Yes/No").

The association of patient socioeconomic status (as measured by part 3, question 6 of the patient protocol survey) and patient reported financial difficulty (as measured by item #28 of the EORTC QLQ-C30) will be tested using multinomial logistic regression modeling. In this model the outcome will be the patient report of financial difficulty, and the covariates will be the patient reported total household income. Item #28 on the EORTC QLQ-C30 is measured as "Not at all/A little/Quite a bit/Very much", converted to 0, 1, 2 and 3 respectively for the purposes of analysis. Part 3, question 6 of the patient protocol survey asks the patient to state the total income of the household they live in and is measured with the following options: less than $20,000, $20,000 - 39,999, $40,000 - 59,999, $60,000 - 79,999, $80,000 - 99,999, $100,000 or more, Don't know.

Descriptive statistics will be used to describe the site-reported plans of developing psychosocial (question 46 of the protocol site survey), transportation (question 33 of the protocol site survey), and financial navigation services (question 24 of the protocol site survey). Each of these questions asks the site to state if they have plans to develop or enhance psychosocial, transportation or financial navigation services (all answered as "Yes/No" questions).

Patterns of financial difficulty (item #28 of the EORTC QLQ-C30) will be assessed utilizing measures of patient reported difficulties paying medical bills (part 1, question 1), delays or foregoing treatment (as measured by part 1, question 4), difficulties covering non-medical expenses due to costs of treatment (part 1, question 7), and financial worries (part 1,). Exploratory LCA will be conducted to assess these patterns. Question 1 asks the participant if they had any problems paying any medical bills in the last twelve months (Yes/No). Question 4 asks the participant if they y delayed medical care because they were worried about the cost (Yes/No). Question 7asks the participant if they had to make any sacrifices in the past 5 years because of debt related to medical care (Yes/No). Question 9 asks the participant if they get sicker or have an accident, how worried are they that they will not be able to pay for their medical bills (Very, Somewhat or Not worried).

This analysis will identify the relationship between distinct patterns of financial burden (as measured by part 1, question 2 of the protocol patient survey) with patient report of financial difficulty (as measured by item #28 of the EORTC QLQ-C30). Financial difficulty will be modeled as a latent class predictor within a multinomial logistic regression in addition to the original LCA measurement model. Item #28 on the EORTC QLQ-C30 is measured as "Not at all/A little/Quite a bit/Very much", converted to 0, 1, 2 and 3 respectively for the purposes of analysis. Part 1, question 2 of the protocol patient survey asks the participant if they or anyone in their family have medical bills that they are unable to pay at all.

The proportion of patients who report receiving financial support (patient survey part 2 question 3 [yes/no]) in the past 12 months will be described using summary statistics.

Logistic regression with receipt of financial support in the past 12 months (patient survey part 2 question 3 [yes/no]) as the dependent variable will be used to test the hypothesis that individuals with multiple myeloma (MM) or chronic lymphocytic leukemia (CLL) will be more likely to report receipt of financial support if they are treated at practices reporting that they offer patients financial guidance through navigators or social workers compared to those treated at practices without these resources. The model will control for disease/treatment characteristics and indicators of patient socioeconomic status.

Aggregate scores of the Valuing Dimensions of the Patient Experience Questionnaire measuring patient concerns regarding treatment and costs of care (part 1, questions 22-36 of the protocol patient survey) will be analyzed descriptively. These questions are all assessed as "Very worried", "Somewhat worried" and "Not worried" will be scored as 2, 1 and 0 respectively and aggregated using standard practices.

Aggregate scores of the Valuing Dimensions of the Patient Experience Questionnaire measuring patient concerns regarding treatment and costs of care (part 1, questions 22-36 of the protocol patient survey) will be stratified by demographics (e.g. gender, race/ethnicity), socioeconomic characteristics (e.g. education, income), disease characteristics (e.g. MM, CLL), and practice-specific factors (e.g. presence of a social worker/patient navigator). These questions are all assessed as "Very worried", "Somewhat worried" and "Not worried" will be scored as 2, 1 and 0 respectively and aggregated using standard practices. Differences in the aggregate scores between groups will be tested using t-tests. Adjustment for multiple comparisons will not be conducted, which is consistent with social science and/or preference based research.

Aggregate scores of the PROMIS-10, EQ-5D-5L and the BAI will be utilized as covariates in a logistic regression model to assess if there is any association with them and the outcome of patient-reported financial difficulty (item #28 of the EORTC QLQ-C30). The PROMIS-10 is a 10-item instrument scored on a scale from 0 - 50 (higher scores = improved QOL). The EQ-5D-5L is a 5-item instrument (5 point scales scored 1-5, where higher scores = worse QOL) and the EQ Visual Analogue scale to score the patient's self-rated health on a scale from 0-100, where higher scores = better QOL. The BAI is a 23-item questionnaire asking patients to rate how often they thought about a variety of topics (Always, Often, Sometimes, Rarely, Never, Not Applicable and Refused to answer). Questions are scored on a scale from 1-5 (Not Applicable and Refused to answer are 88 and 99, respectively, and excluded from scoring), and the questionnaire is scored using standard scoring algorithms.