Translation, cultural adaptation and validation of the Danish version of the haematological malignancy patient-reported outcome measure (HM-PRO).

Introduction

Assessment of cancer patients’ quality of life (QoL) through patient-reported outcomes (PRO) during and after treatment is gaining ground. In fact, implementation of PROs into clinical practice is a strategic target of leading cancer organizations both in Europe and in the United States [ 1 ]. Patients with haematological malignancies will often be highly affected by disease symptoms and/or side effects to treatment. To capture this, systematic use of PROs may help to better understand the actual symptom and treatment burden experienced by patients themselves. Further, the use of PROs in routine clinical practice may be highly beneficial for improving patient-physician communication and to prioritize patient-centred care [ 2 , 3 ], thereby improving the QoL in patients affected by haematological malignancies. In Europe, one of the most widely used cancer PRO instrument is the EORTC QLQ-C30 questionnaire, which is regarded as a cancer generic questionnaire for use in solid cancer clinical trials [ 4 ]. Also, the EORTC QLQ-C30 has been used as a treatment effect measure in several haematological trials [ 5 ]. However, recent studies in patients with haematological malignancies have shown that some symptoms are unique to this subgroup of cancer patients, and may thus not be captured by the existing generic cancer questionnaires, originally designed for use in solid cancers [ 6 , 7 ]. Therefore, the “Scientific Working Group for Quality of Life and Symptoms”, within the European Haematology Association, decided to develop a new generic PRO designed specifically for and by patients with haematological malignancies (the HM-PRO) [ 8 , 9 ]. With increasing importance of measuring QoL in patients with haematological malignancies including issues that are important to patients, a new validated questionnaire could be an important instrument for both clinicians and patients in daily clinical practice as well as in clinical trials.

The HM-PRO consists of 42-items, divided into two parts (A and B). Part A comprises 24 items covering daily impacts on patients´ QoL. It is grouped into four subdomains; Physical well-being (7 items), Social well-being (3 items), Emotional behaviour (11 items), and Eating and Drinking (3 items). The recall period for part A is “today”. Part B consists of 18 items and constitutes a single domain targeting patient-experienced disease symptoms and treatment side effects specifically related to haematological malignancies. The recall period for part B is the “past 3 days”. Items in Part A and B are rated on a 3-point Likert scale ranging from “not at all” (0 point) to “a lot” (2 points). In addition, a “not applicable” (0 point) response option is available too in Part A and is incorporated in the calculation of the subdomain scores in Part A. Subdomain scores are standardized to a 0 to 100-point scale, with higher scores meaning greater impact on QoL and a higher symptom burden. No total HM-PRO score is calculated [ 10 ].

Within the past two decades, the Danish Health Authority has launched five national cancer plans with the aim of improving cancer treatment across all diagnoses. In the fourth cancer plan (Cancer Treatment Package IV) from 2016 ( www.sst.dk ), it is specified that the use of PROs in clinical care should be optimized. The intention is to systematically collect PRO data on health-related QoL in Danish cancer patients, including those affected by haematological malignancies. It is an expressed goal to further incorporate PRO data into the National Danish Clinical Registries, and therefore, a validated robust and adequate generic PRO for patients with haematological malignancies is highly needed.

Advances in the treatment of haematological cancers have led to improvement in survival putting emphasis on the long-term effects of the disease and its treatments affecting the health‐related quality of life (HRQoL) of HM patients both short term and long term. Difficulty with physical and psychosocial activity, living with uncertainty, worrying about future and relapse and impact on work life are evident in survivors of HM and bone marrow transplant.

A comprehensive list of quality-of‐life issues important to patients suffering from haematological malignancies was presented in a systemic review and a list of HRQoL instruments used in haematological malignancies in both daily clinical practice and research provided [ 6 ]. However, none of the instruments reviewed (except MyPOS) appeared to have been developed specifically for patients with HM and for use in clinical practice. Evidence suggests that the psyche of patients with HMs is different to that of patients with solid tumours, and the conceptual model for the two is quite different. MH patients experience significant challenges of emotional/psychological disturbances, prolonged periods of neutropenia, more frequent visits to intensive care units, are more likely to die in hospital, receive less information on the sexual side effects of treatment and are much less likely to be referred to specialist palliative care at the end of their illness [ 6 ].

The aim of this study was (1) to translate and cross-culturally adapt the HM-PRO into Danish and (2) to psychometrically evaluate its performance in Danish patients with haematological malignancies.

Materials and methods

Study design

This study was a nationwide project with participation from six haematology departments across Denmark sponsored by the Department of Haematology, Odense University Hospital, Denmark. The study included three steps; (1) translation and cross-cultural adaptation, (2) cognitive debriefing and (3) psychometric testing. According to Danish law, this study was exempted for notification to the regulatory authorities. The study was approved by the Danish Data Protection Agency (Journal number 19/13870).

Translation, cross-cultural adaptation and cognitive debriefing

Translation and cross-cultural adaptation of the original English HM-PRO into Danish was conducted and followed standardised guidelines for forward and backward translation procedures [ 11 – 14 ] (detailed methodology is given in the supplementary materials). Cognitive debriefing interviews were conducted on 11 patients.

Psychometric testing of the Danish version of the HM-PRO

Subsequently, the approved Danish version of the HM-PRO was empirically tested for structural validity to confirm a number of constructs of the original English HM-PRO. For this, participants were recruited consecutively at the aforementioned six departments and asked to complete the Danish HM-PRO once. Both inpatients and outpatients were eligible for participation. Besides being diagnosed with one of the haematological malignancies as listed below the inclusion criteria for participation were: 18 years or older, able to understand and read Danish, and sufficient cognitive function (subjectively decided by a responsible healthcare professional). Participants were excluded if they had known comorbidities resulting in a high level of symptoms not related to their haematological disease and/or if they had an expected remaining life of less than 3 months from study enrolment.

To achieve a wide range of the most common haematological diagnoses in accordance with the generic characteristics and constructs of the HM-PRO, eligible participants followed the 2008 WHO classification of tumours of haematopoietic and lymphoid tissues [ 15 ], being:

Acute myeloid leukaemia and related precursor neoplasms including therapy-related myeloid neoplasms and myeloid sarcoma (AML) Acute lymphoid leukaemia) (ALL) Chronic lymphocytic leukaemia (CLL) Chronic myeloid leukaemia (CML) Multiple myeloma (MM) Aggresive non-Hodgkin’s lymphoma (ANHL) Indolent non-Hodgkin’s lymphoma (INHL) Hodgkin’s lymphoma (HL) Myeloproliferative neoplasm (MPN) Myelodysplastic (MDS)

Data collection procedure

At each of the six participating departments, a site investigator was responsible for the recruitment of participants. Eligible participants were provided with oral and written study information. Patients who accepted to participate signed an electronic consent form prior to data collection. To ensure an equal distribution of participants in accordance with the ten haematological sub diagnoses, the overall participant characteristics were monitored as the process of inclusion progressed and gave input to the site investigators about the current recruitment status in terms of diagnoses and sociodemographic variation. As such, the site investigators were able to target their recruitment of participants in accordance with the recruitment strategy.

Sociodemographic characteristics and clinical data of the participants were collected via self-report as well as physician-report. Self-reported data using a tablet concerned the following; age, gender, education (highest grade completed), and known comorbidity. Physician-reported data, extracted from medical records, concerned the following; cancer type, time since diagnosis, treatment history, current treatment, and ECOG performance status.

Data management

Data were stored in REDCap, an electronic web-based data capture system, hosted by Open Patient data Exploratory Network (OPEN) at Odense University Hospital, Denmark. The REDCap system managing the electronic data collection was designed so that participants were required to answer all questions, thus resulting in no missing data. Patients who were unwilling/unable to complete the HM-PRO electronically were allowed to self-complete the survey using hard copy.

Statistical analyses

Descriptive statistics included mean and standard deviation (SD). Correlation and regression coefficient for each item were presented. All statistical analyses were performed using STATA 15.0 software, with probability of Type I error set at p < 0.05 level.

Validity

Validity was measured by construct validity, which is defined as the degree to which the scores of a measurement instrument are an adequate reflection of the dimensionality of the construct to be measured. As exploratory analysis of the factor structure and number of dimensions/scales was performed during the development of the original HM-PRO [ 9 ], confirmatory factor analysis (CFA) was used in this study to examine whether the data fit the predetermined model [ 16 ]. No listwise or pairwise deletion was carried out. Instead, a likelihood ratio method, e.g. full information maximum likelihood (FIML) or expectation-maximization algorithm (EM) was used to calculate estimates on all available data.

According to COSMIN, the minimum requirement of a sample size to conduct CFA is 7 subjects per item [ 17 , 18 ]. As the HM-PRO is a 42-item questionnaire, a sample size of 294 participants was thus the target for this CFA. To comply with the intention of an equal distribution of haematological malignancies, attempts to recruit 42 patients from each of the seven haematological subgroups (as described above) were made. This applied also to variation in socio-demographic variables (e.g. age , gender , educational levels ) and disease states (i.e. stable, remission, progression).

The HM-PRO data were not normally distributed, therefore for CFA and bi-factor modelling the Weighted Least Squares with Mean and Variance adjusted (WLSMV) estimator [ 19 , 20 ] was used that is designed for ordinal data, is asymptotically distribution free, and does not make distributional assumptions about the observed variables. Fit parameters scaled for the WLSMV were used to test whether the proposed model was superior to alternative models. CFA and IRT were used to test proposed models and evaluation of model fit was performed using the following fit parameters. Root Mean Square Error of Approximation (RMSEA) expressing the lack of fit per degree of freedom of the model with values interpreted as follows: ≤0.05 = very good; >0.05–0.099 = good; ≥0.10 poor fit. The comparative fit index (CFI) assesses fit relative to a null model and ranges from 0 to 1 with values of 0.90–0.95 indicating acceptable and > 0.95 good fit. Tucker–Lewis index (TLI) adjusts for the number of model parameters and is interpreted as for CFI. The Standardised Root Mean Square Residual (SRMR) including 90% confidence intervals (CI) is the average of the differences between the observed and predicted correlations and has a range from 0 to 1 with values < 0.08 indicating good fit. Internal consistency was determined by Cronbach’s alpha [ 21 , 22 ] for Part A (and its domains) calculated using IBM SPSS Statistics version 27.

Bi-factor analysis was performed using R version 4.2.2 (R Foundation for Statistical Computing) with package Lavaan version 0.6–16. The estimator used was maximum likelihood (ML) and missing values were handled with full information maximum likelihood (FIML) estimation for latent variables. The Omega function in Lavaan was used to guide selection of factors for bi-factor modelling, additionally using factor loading.

Differential Item Functioning (DIF) analyses were performed across two variables: gender and age. An age cut-off at 69 years (i.e. median) was used to provide balanced groups, and male vs. female was used as per the data. Item response theory for DIF was performed by IRTPRO version 4.2 (Scientific Software International Inc., Chapel Hill, NC, USA), using a Graded Response Model (GRM) with 3 categories. Independent models were run for Part A (24 items) and Part B (18 items).

Results

Translation and cross-cultural adaptation

The forward and backward translations did not provide any major issues. However, minor issues, mostly related to cross-cultural matters, were solved through consensus with participation from both forward and backward translators. e.g., the item “I feel distressed” had to be translated into Danish with the use of adjectives (Jeg er frustreret/ramt/ulykkelig), as the Danish language do not have a single word that covers all the constructs of being distressed. Thus, the Danish version of this item contained three adjectives separated with slashes, rather than commas, to ensure that patients did not have to feel affected by all three aspects as at the same time to be able to correctly rate their level of distress.

Cognitive debriefing

The cognitive debriefing interviews were conducted on 11 patients in total (4 females; 3 newly diagnosed, 7 stable/in remission and 1 in relapse; 4 aged < 50, 6 aged between 50 and 80 and 1 aged > 80; and 3, 2 and 6 with low, mid and high educational status, respectively. The diagnoses were as follows: HL = 2, MDS = 2, MPN = 1, AML = 3, MM = 1, and DLBCL = 1.

Only item number 7 in the emotional domain ( I do not feel confident ) needed careful consideration and discussion. Three of eleven patients mentioned this item to be confusing and difficult to answer. This was due to the item being formulated differently compared to all the other items. For example, it is phrased in a reversed manner to comply with the fact that all the items of the HM-PRO concerns impairment and, as such, this item uses negative connation. This created doubts on how to answer the question correctly. Thus, to emphasize that this item needs special attention before answering (e.g. due the negative wording) a note in brackets was added to this item in the Danish version and now reads “Jeg føler mig ikke selvsikker (vær opmærksom på, at der står IKKE)”. This revised Danish version of this item was tested on four patients and no further concerns were raised. A final version of the Danish HM-PRO was approved in August 2019.

Psychometric validation

Patient characteristics

From August 2019 to June 2020, 295 participants signed up for inclusion for the psychometric validation of the Danish HM-PRO. Recruitment of patients from the six departments was distributed as follows: Aalborg 70, Aarhus 49, Odense 39, Roskilde 28, Herlev 71, and Copenhagen University Hospital 44 patients. However, despite signed consent, the HM-PRO was left empty in one case, and in another case the HM-PRO was only partly completed. The sample for complete-case analyses thus constituted 293 patients. The majority was outpatient, constituting 75% of the total sample (222/293). Mean age was 65 years (SD 15) and 137 (47%) were females (Table 1 ). The largest hematologic group was patients with aggressive non-Hodgkin’s lymphoma constituting 17% of the total sample (51/293) (Table 1 ). A certain number of study participants requiring assistance with completing any PRO instrument is not unusual. With reference to this study in particular, there were some people with limited visual ability who needed assistance to complete questionnaires or who needed additional explanation for using an electronic device (Table 1 ), rather than on the questionnaire itself, but this did not affect validity testing of the Danish HM-PRO in any way.

Table 1 Demographic data and clinical characteristics ( n = 293) n (%) Age , mean (SD) 64.8 (15) Median 69 Range 19–92 Sex -Female 137 (47) -Male 156 (53) Setting (at questionnaire completion) -Inpatient 71 (25) -Outpatient 222 (75) Educational level -Elementary school 47 (16) -Vocational basic course 110 (38) -Medium cycle higher education (3–4 years) 87 (30) -Long cycle higher education (> 4 years) 49 (17) Civil status - Unemployed (looking for job) 5 (2) -Employed (full-time) 47 (16) -Employed (part-time) 15 (5) -Student 6 (2) -Retired 187 (64) -On sick leave 33 (11) Haematological diagnosis - Acute myeloid leukaemia 27 (9) - Acute lymphoid leukaemia 14 (5) - Chronic lymphocytic leukaemia 32 (11) - Chronic myeloid leukaemia 17 (6) - Multiple myeloma 44 (15) - Aggresive non-Hodgkin’s lymphoma 51 (17) - Indolent non-Hodgkin’s lymphoma 31 (11) - Hodgkin’s lymphomah 15 (5) - Myeloproliferative neoplasm 34 (12) - Myelodysplastic syndrome 26 (9) ECOG (for the outpatient patients only) -0 98 (44) -1 96 (43) -2 19 (9) -3 6 (3) -4 1 (< 1) Missing 1 (< 1) Time since diagnose (months) , mean (SD) ≤ 6 months 78 (27) 6–12 months 32 (11) 12–36 months 58 (20) > 36 months 120 (41) Missing 5 (2) Disease status -Stable 118 (40) -Remission 74 (25) -Progression 93 (32) Missing 8 (3) In active treatment (at questionnaire completion) -yes 205 (71) -no 84 (29) If yes, which treatment: - Allogenic transplantation 2 (1) - Autologous transplantation 3 (2) - Chemotherapy 149 (73) - Non-chemotherapy 41 (20) - Other 10 (4) HM-PRO completion - Hard copy 32 (11) - Electronic 261 (89) - Needed assistance with completion of the electronic version 64 (22) - Needed assistance with reading out loud the questions 37 (13)

Validity

Internal consistency

Internal consistency as calculated by Cronbach’s alpha is shown in Table 2 .

Table 2 Internal consistency of the HM-PRO HM-PRO HM-PRO domains This study (Danish) Goswami 1 Revised version 2 Assessment 1 (t0) 2 Assessment 2 (t7) 2 Part A Overall Part A (24 items) 0.81 0.95 0.91 0.90 0.90 PB = Physical behaviour (7 items) 0.66 0.88 0.91 SW = Social wellbeing (3 items) 0.46 0.70 0.68 ES = Emotional state (11 items) 0.72 0.85 0.87 ED = Eating/Drinking (3 items) 0.78 0.84 0.85 Part B All items (18 items) 0.84 0.85 0.76 0.80 0.80 1 Final revised version: Goswami et al. [ 9 ]. 2 Development versions: Goswani [ 23 ]

Structural validity

There was no missing data in the dataset used for psychometric analysis. The CFA of the original 4-factor model for Part A of the HM-PRO revealed a RMSEA of 0.077 (90%CI 0.070–0.084), a SRMR value of 0.095, and a CFI and TLI of 0.92 and 0.91, respectively (Table 3 ). This model was superior to two explorative 3-factor models where the physical and social domains, and the social and emotional domains, respectively were combined into one domain (Table 3 ). The CFA of the 1-factor model for Part B of the HM-PRO revealed a RMSEA of 0.06 (90%CI 0.05–0.07 ), and a CFI and TLI of 0.94 and 0.93, respectively (Table 3 ).

Table 3 The model statistics of part A and part B of the Danish version of the HM-PRO Model Part A X 2 P CFI TLI RMSEA (90% CI) SRMR 4 factor model (the original model) - Complete case analysis 623.45 < 0.001 0.922 0.912 0.077 (0.070–0.084) 0.095 3 factor model - Physical and Social domains combined 726.58 < 0.001 0.910 0.901 0.082 (0.075–0.088) 0.105 3 factor model - Social and emotional domains combined 695.14 < 0.001 0.913 0.903 0.080 (-0.74-0.087) 0.100 Model Part B X 2 P CFI TLI RMSEA (90% CI) SRMR 1 factor model (the original model) - Including all patients that have answered on just some of the questionnaire 241.99 < 0.001 0.936 0.928 0.060 (0.050–0.070) 0.095

Bi-factor analyses

The bi-factor of the original 4-factor model for Part A revealed a RMSEA of 0.078 (90%CI 0.070–0.084), a SRMR value of 0.095, and a CFI and TLI of 0.92 and 0. 91, respectively (Table 4 ). The fit of this model was marginally better than the 4-factor CFA model. Bi-factor models based on factors taken from hierarchical models (Lavaan Omega, see supplementary Figs. 1–3) are also shown in Table 4 . The 4-factor model based on the hierarchal model gave better fit than the original 4-factor model. It was not appropriate to fit a bi-factor model to Part B as it has been shown to be unidimensional.

Table 4 The model statistics of part A bifactor of the Danish version of the HM-PRO Model X 2 P CFI TLI RMSEA (95% CI) SRMR 4 factor bifactor model (the original model) 1 623.45 < 0.001 0.922 0.912 0.077 (0.070–0.084) 0.095 2 factor bifactor model from Omega 2 304.81 < 0.001 0.962 0.953 0.056 (0.048–0.064) 0.071 3 factor bifactor model from Omega 3 265.69 < 0.001 0.971 0.965 0.048 (0.040–0.057) 0.066 4 factor bifactor model from Omega 4 182.77 < 0.001 0.983 0.978 0.038 (0.028–0.047) 0.054 PB = Physical, behaviour, SW = Social wellbeing, ES = Emotional state, ED = Eating/drinking 1 factor1 = PB1+PB2+PB3+PB4+PB5+PB6+PB7, factor2 = SW1+SW2+SW3, factor3 = EB1+EB2+EB3+EB4+EB5+EB6+;EB7+EB8+EB9+EB10+EB11, factor4 = ED1+ED2+ED3 2 factor1 = PB1 +PB2+PB3+PB4+PB5+PB6+PB7+ED1+ED2+ED3+SW1+EB1+EB9+EB10, factor2 = EB1+EB2+EB3+EB4+EB5+EB6+EB7+EB8+EB9+EB10+EB11+SW1+SW2+PB6 3 factor1 = EB1+EB2+EB3+EB4+EB5+EB6+EB7+EB8+EB9+EB10+EB11+SW2+SW3, factor2 = PB1+PB2+PB3+PB4+PB5+PB6+PB7+SW1, factor3 = ED1+ED2+ED3 4 factor1 PB6+EB1+EB2+EB3+EB7+EB10+SW1+SW2+SW3, factor2 = PB1+PB2+PB3+PB4+PB5+PB7, factor3 = EB4+EB5+EB6+EB8+EB11, factor4 = ED1+ED2+ED3+EB9

DIF analyses

Item response theory with a dataset N = 293 was used to investigate DIF. All factor loadings > 0.36 and almost all local dependency (LD) statistic < 10 indicated minimal local dependency [ 24 ]. For Part A, S-X 2 item level diagnostic statistics indicated optimal fit for items with p > 0.05 for all items and a RMSEA of 0.07. Item trace lines indicated misfitting of items PB6, PB7, and SW3, attributable to missing data. For Part B, S-X 2 item level diagnostic statistics indicated a suboptimal fit for items 3 (SS3) only, with p < 0.05 and a RMSEA of 0.04. Item trace lines indicated misfitting of item SW 4 only.

The DIF by gender on Part A was performed with Group 1 (male) = 155 and Group 2 (female) = 137. All factor loadings in both groups > 0.42. From S-X 2 item level diagnostic statistics only item 12 (EB2) in both groups had suboptimal fit with p < 0.05. The DIF analysis by X 2 testing showed overall DIF in items 10 (SW3), 13 (EB3) and 16 (EB6) due location (uniform) DIF ( p < 0.05). RMSEA was 0.0 (Table 5 ).

DIF by age for Part A with Group 1 = 145 (age < 69) Group2 = 147 (age > = 69) showed uniform DIF in Items 1 (PB1), 2 (PB2), 5 (PB5), 12 (EB2) and 16 (EB6), discrimination/non-uniform DIF in items 10 (SW3), 13 (EB3), 20 (EB10), and 23 (ED2), and both uniform and non-uniform DIF in items 6 (PB6) and 22 (ED1) (Table 5 ).

Table 5 DIF statistics for graded items part A by gender and age DIF statistics for graded items Part A by gender Item Group 1 Group 2 Total X 2 d.f. p X 2a d.f. p X 2c/a d.f. p SW3 10 10 24.6 3 0.0001 1.7 1 0.1877 22.8 2 0.0001 EB3 13 13 18.5 3 0.0003 1.4 1 0.2353 17.1 2 0.0002 EB6 16 16 11.5 3 0.0091 1.4 1 0.2352 10.1 2 0.0063 DIF statistics for graded items Part A by age Item Group 1 Group 2 Total X 2 d.f. p X 2a d.f. p X 2c/a d.f. p SS1 1 1 8.2 3 0.0422 0.0 1 0.9508 8.2 2 0.0167 SS3 3 3 10.0 3 0.0186 0.0 1 0.8883 10.0 2 0.0068 SS8 8 8 9.9 3 0.0194 3.2 1 0.0723 6.7 2 0.0356 Total χ2 is overall DIF, χ2a for slope (discrimination, non-uniform), χ2c/a for location (uniform)

DIF by gender on Part B was performed with the same grouping and showed no DIF in any items. RMSEA was 0.03. DIF by age for Part B with the same grouping showed uniform DIF for items 1 (SS1), 3 (SS3), and 8 (SS8). RMSEA was 0.04 (Table 6 ).

Table 6 DIF statistics for graded items part B by age Item Group 1 Group 2 Total X 2 d.f. p X 2a d.f. p X 2c/a d.f. p SS1 1 1 8.2 3 0.0422 0.0 1 0.9508 8.2 2 0.0167 SS3 3 3 10.0 3 0.0186 0.0 1 0.8883 10.0 2 0.0068 SS8 8 8 9.9 3 0.0194 3.2 1 0.0723 6.7 2 0.0356

Discussion

In the present study, we have translated and cross-culturally adapted the HM-PRO into Danish and tested its psychometric properties in Danish patients with a variety of haematologic malignancies while retaining the same factor structure as the original version of the HM-PRO. Few issues were met with regards to the translation and adaptation of the HM-PRO into Danish. After cognitive debriefing, a revised version of this item was tested on four patients. No further concerns were raised and the final version of the Danish HM-PRO was approved.

Confirmatory factor analysis

The CFA for both parts of the HM-PRO provided evidence to support the suggested factor structure of the HM-PRO. In the present study, the RMSEA was within the acceptable levels of error of approximation, while the SRMR value was slightly below the cut-off, thus not indicating a poor fit. With regards to the CFI and TLF values, these were slightly outside the suggested range, but within acceptable distance to support the model fit. This is based on the fact that the guidelines for acceptable model varies and as the sample size was at the lower end of the suggested numbers for CFA. Thus, future studies should aim to include even larger sample sizes.

Internal consistency as calculated by Cronbach’s alpha was 0.81 for Part A and 0.84 for Part B of the Danish HM-PRO confirming the internal consistency of both scales [ 25 ]. It should be noted that it is expected that the magnitude of Cronbach’s alpha is related to the number of items used in its calculation, and will therefore increase as the number of items in the calculation increases [ 17 ]. This effect often tends to make domains with few items such as the social/wellbeing and eating/drinking with three items each to “underperform” and those with many items “overperform”. However, these items came from patient consultation with 130 patients, and when the item list was saturated and duplicate items removed, those were clearly related conceptually combined into their own domains and validated, first by EFA then by CFA, so these domains with few items are a real construct in capturing QoL of haematological malignancies.

BIfactor and DIF analysis

Bifactor analysis for Part A of the HM-PRO showed slightly better fit for the original 4-factor model than the CFA analysis, and a slightly modified 4-factor bifactor model based on a hierarchical model improved this further. Interestingly, the hierarchical model which is completely unspecified except for the number of factors to obtain, using only the patient responses with no predetermined relationships, mapped very closely to the original domains hypothesised during development [ 10 ], and confirmed by CFA.

Generally, items showed good fit in the IRT models with minimal local dependency. The DIF analysis showed limited differential item functioning by gender. This would be expected as this is a baseline study of patients already diagnosed with haematological malignancy (so incidence, prognosis, response to treatment and survival are not relevant). Sztankay et al. reviewed gender differences in QoL in patients with haematological malignancies [ 26 ] and found female gender proved to be one predictor of impaired overall QoL in lymphoma patients, while representing the only variable associated with lower global QoL in a sample with chronic lymphocytic leukaemia (CLL). Diminished functioning and poor adjustment in occupational and social domains are equally well documented in female patients with HL or CLL. Previous study has explored the impact of gender on distress, anxiety, and depression in haematological cancer patients receiving chemotherapy [ 27 ] and found significant interaction between gender and practical and physical problems. The DIF by gender might be attributable to differences in perception of pain [ 28 ] and the increased risk of depression in women [ 29 ]. Both of these factors are significant predictors of QoL in female cancer patients [ 30 ].

However, when DIF analysis was performed on groups by age, significant DIF, both non-uniform and uniform, was revealed. A study has previously shown that age has a significant negative correlation with global QoL, physical functioning and role functioning (using EQ-5D and EORTC QLQ‐C30 instruments) as well as increasing age was also associated with lowest QoL [ 31 ]. Priscilla et al. [ 32 ] reported similar findings of reduced QoL with impaired physical and role functioning in older patients with a haematological cancer, and the trend reported by Slovacek et al. [ 33 ]

Strengths

A strict translation guideline was followed including independent forward and backward translations [ 14 ]. This nationwide study with inclusion of patients with a variety of hematologic malignancies, sex, age, educational level, etc.) is a representative population for Danish patients with hematologic malignancies, thus improving the generalisability of the study. Furthermore, rigorous psychometric testing applied throughout the development of the Danish HM-PRO should provide the clinicians as well as the researchers with full confidence in the reliability and validity of the generated scores and its application to aid treatment decision-making.

Limitations

The sample size for this study was smaller than desired for the analyses performed resulting in the fits for CFA (RMSEA, CLI and TLI) slightly below desired cutoffs. Also, although responsiveness values as well as other measures of validity, score bands systems and ‘Minimum important difference’ (MID) are available from studies of the source HM-PRO version, these were not assessed in the present study. As psychometric properties, in general, are context and population dependent, these values should also be investigated in a Danish context. Although equivalency was shown between electronic and paper versions during psychometric testing of the original HM-PRO [ 34 ], this should be further explored in the Danish version. Moving forward with the use of the Danish HM-PRO, there will be a need to test its responsiveness or longitudinal application with appropriately designed studies and these are being currently conducted within Denmark.

Conclusions

The HM-PRO is the first generic haematological specific PRO measure for use in clinical practice and clinical trials. Such a generic tool for health-related QoL measurement is needed in Denmark for several reasons. Among others, The Danish Government and Danish Regions aim to use PROs broadly within the Danish healthcare system: First, to increase and improve symptom monitoring during and after cancer treatment; and second, to optimize patient engagement during routine cancer care. One rationale for this is that routine use of PRO data has been found to clinically benefit patients during cancer treatment [ 35 , 36 ].

However, since health-related QoL of patients with haematological malignancies differ from patients with solid tumours, there is a need for specific and generic tools developed specifically for haematological patients [ 6 ]. Feasibility in PRO data collection in daily clinical practice and minimization of patient burden has to be considered before implementation of PRO data in provision of healthcare. Therefore, a PRO instrument that can be used to monitor symptoms during routine cancer care as well as quality assessment of treatment through The Danish Clinical Registries will be of utmost importance.

The HM-PRO has demonstrated favourable psychometric properties for use in clinical practice, symptoms monitoring in registries as well as in clinical trials. Although the HM-PRO is a relatively new instrument, it has already undergone cross-cultural translation and been adapted into at least 22 other languages ( https://hmpro.co.uk ).

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Change history

5/22/2025

For author Lene Kongsgaard Nielsen affiliation 5 was replace with affiliation 10.

Acknowledgements

Our heartfelt thanks go to Henrik Rode Eshøj. Sadly, Henrik passed away during the time of analyzing the data and writing the manuscript. Henrik made a great contribution as project manager of the study. He was a key manager in writing the protocol and statistical analysis plan. Also, together with the local investigators at each site, Henrik organized the data collection and supported the study coordinators. In addition, Henrik performed a great part of the analysis before he passed away. Special thanks go to Dr Jeff Johns for performing additional analysis, and to the entire study team and the patients for making the study possible. The study was funded by Department of Haematology, Odense University Hospital, Denmark.

Author contributions

SS, LKN and Henrik Rode Eshøj initiated the study in Denmark. Every author collaborated with conceptualisation and protocol writing of the study. Henrik Rode Eshøj was the daily project leader on behalf of the study sponsor until he passed away and thereafter TR took over. MTL, PA, HG, MTL, NA, KKP and CLA initiated the study at the participating sites, recruited patients and collected data. SM, Henrik Rode Eshøj, SS and Jeffrey Johns performed the analyses and interpreted the results. SS, LKN, TR, KKP, Henrik Rode Eshøj and Jeffery Johns wrote the first draft of the manuscript. All authors reviewed the manuscript and approved the final version before submission.

Funding

No external funding was used for this study.

Data availability

Data available on request due to privacy/ethical restrictions.

Declarations

Ethical approval

According to Danish law, this study was exempted for notification to the regulatory authorities. The study was approved by the Danish Data Protection Agency (Journal number 19/13870).

Consent for publication

All the authors have given consent of publication of this manuscript in The Journal of Patient Reported Outcomes.

Conflict of interest

SS is a joint copyright holder of the HM-PRO. LKN, TR, CLA, PA, MTL, SM, KKP, MTL, NA have no competing interests.

References

1. Efficace F Gaidano G Lo-Coco F Patient-reported outcomes in hematology: is it time to focus more on them in clinical trials and hematology practice? Blood 2017 130 7 859 866 10.1182/blood-2017-03-737403 28694324 Efficace F, Gaidano G, Lo-Coco F (2017) Patient-reported outcomes in hematology: Is it time to focus more on them in clinical trials and hematology practice? Blood 130(7):859–866 28694324 10.1182/blood-2017-03-737403

2. Bredart A Bottomley A Blazeby JM Conroy T Coens C D’Haese S An international prospective study of the EORTC cancer in-patient satisfaction with care measure (EORTC IN-PATSAT32) Eur J Cancer 2005 41 14 2120 2131 10.1016/j.ejca.2005.04.041 16182120 Bredart A, Bottomley A, Blazeby JM, Conroy T, Coens C, D’Haese S et al (2005) An international prospective study of the EORTC cancer in-patient satisfaction with care measure (EORTC IN-PATSAT32). Eur J Cancer 41(14):2120–2131 16182120 10.1016/j.ejca.2005.04.041

3. Snyder CF Aaronson NK Choucair AK Elliott TE Greenhalgh J Halyard MY Implementing patient-reported outcomes assessment in clinical practice: a review of the options and considerations Qual Life Res 2012 21 8 1305 1314 10.1007/s11136-011-0054-x 22048932 Snyder CF, Aaronson NK, Choucair AK, Elliott TE, Greenhalgh J, Halyard MY et al (2012) Implementing patient-reported outcomes assessment in clinical practice: A review of the options and considerations. Qual Life Res 21(8):1305–1314 22048932 10.1007/s11136-011-0054-x

4. Aaronson NK Ahmedzai S Bergman B Bullinger M Cull A Duez NJ The European organization for research and treatment of cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology J Natl Cancer Inst 1993 85 5 365 376 10.1093/jnci/85.5.365 8433390 Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ et al (1993) The European organization for research and treatment of cancer QLQ-C30: A quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 85(5):365–376 8433390 10.1093/jnci/85.5.365

5. Nielsen LK Jarden M Andersen CL Frederiksen H Abildgaard N A systematic review of health-related quality of life in longitudinal studies of myeloma patients Eur J Haematol 2017 99 1 3 17 10.1111/ejh.12882 28322018 Nielsen LK, Jarden M, Andersen CL, Frederiksen H, Abildgaard N (2017) A systematic review of health-related quality of life in longitudinal studies of myeloma patients. Eur J Haematol 99(1):3–17 28322018 10.1111/ejh.12882

6. Goswami P Khatib Y Salek S Haematological malignancy: are we measuring what is important to patients? A systematic review of quality-of-life instruments Eur J Haematol 2019 102 4 279 311 10.1111/ejh.13203 30556217 Goswami P, Khatib Y, Salek S (2019) Haematological malignancy: Are we measuring what is important to patients? A systematic review of quality-of-life instruments. Eur J Haematol 102(4):279–311 30556217 10.1111/ejh.13203

7. Salek S Kamudoni P Oliva E Ionova T Quality of life issues important to patients with haematological malignancies Value Health 2015 18 7 A709 10.1016/j.jval.2015.09.2668 Salek S, Kamudoni P, Oliva E, Ionova T (2015) Quality of life issues important to patients with haematological malignancies. Value Health 18(7):A709

8. Goswami P Oliva EN Ionova T Else R Kell J Fielding AK Hematological malignancy specific Patient-Reported outcome measure (HM-PRO): construct validity study Front Pharmacol 2020 11 1308 10.3389/fphar.2020.01308 33013368 PMC7506039 Goswami P, Oliva EN, Ionova T, Else R, Kell J, Fielding AK et al (2020) Hematological malignancy specific Patient-Reported outcome measure (HM-PRO): Construct validity study. Front Pharmacol 11:1308 33013368 10.3389/fphar.2020.01308 PMC7506039

9. Goswami P Oliva EN Ionova T Else R Kell J Fielding AK Development of a novel hematological malignancy specific Patient-Reported outcome measure (HM-PRO): content validity Front Pharmacol 2020 11 209 10.3389/fphar.2020.00209 32210809 PMC7066982 Goswami P, Oliva EN, Ionova T, Else R, Kell J, Fielding AK et al (2020) Development of a novel hematological malignancy specific patient-reported outcome measure (HM-PRO): Content validity. Front Pharmacol 11:209 32210809 10.3389/fphar.2020.00209 PMC7066982

10. Goswami P, Salek S, Ionova T, Oliva EN, Fielding AK, Karakantza M et al (2017) HM-PRO: A Novel Patient-Reported Outcome Measure in Hematological Malignancy for Use in Clinical Practice. Blood 130(Suppl 1):2176

11. Beaton DE Bombardier C Guillemin F Ferraz MB Guidelines for the process of cross-cultural adaptation of self-report measures Spine (Phila Pa 1976) 2000 25 24 3186 3191 10.1097/00007632-200012150-00014 11124735 Beaton DE, Bombardier C, Guillemin F, Ferraz MB (2000) Guidelines for the process of cross-cultural adaptation of self-report measures. Spine (Phila Pa 1976) 25(24):3186–3191 11124735 10.1097/00007632-200012150-00014

12. Bullinger M Alonso J Apolone G Leplege A Sullivan M Wood-Dauphinee S Translating health status questionnaires and evaluating their quality: the IQOLA project approach. International quality of life assessment J Clin Epidemiol 1998 51 11 913 923 10.1016/S0895-4356(98)00082-1 9817108 Bullinger M, Alonso J, Apolone G, Leplege A, Sullivan M, Wood-Dauphinee S et al (1998) Translating health status questionnaires and evaluating their quality: The IQOLA project approach. International quality of life assessment. J Clin Epidemiol 51(11):913–923 9817108 10.1016/s0895-4356(98)00082-1

13. Guillemin F Bombardier C Beaton D Cross-cultural adaptation of health-related quality of life measures: literature review and proposed guidelines J Clin Epidemiol 1993 46 12 1417 1432 10.1016/0895-4356(93)90142-N 8263569 Guillemin F, Bombardier C, Beaton D (1993) Cross-cultural adaptation of health-related quality of life measures: Literature review and proposed guidelines. J Clin Epidemiol 46(12):1417–1432 8263569 10.1016/0895-4356(93)90142-n

14. Koller M Kantzer V Mear I Zarzar K Martin M Greimel E The process of reconciliation: evaluation of guidelines for translating quality-of-life questionnaires Expert Rev Pharmacoecon Outcomes Res 2012 12 2 189 197 10.1586/erp.11.102 22458620 Koller M, Kantzer V, Mear I, Zarzar K, Martin M, Greimel E et al (2012) The process of reconciliation: Evaluation of guidelines for translating quality-of-life questionnaires. Expert Rev Pharmacoecon Outcomes Res 12(2):189–197 22458620 10.1586/erp.11.102

15. Swerdlow SH Campo E Harris NL WHO classification of tumours of Haematopoietic and lymphoid tissues 2008 Lyon IARC Swerdlow SH, Campo E, Harris NL (2008) WHO classification of tumours of haematopoietic and lymphoid tissues. IARC, Lyon

16. Schreiber JB Nora A Stage FK Barlow EA King J Reporting structural equation modeling and confirmatory factor analysis results: A review J Educ Res 2006 99 6 323 338 10.3200/JOER.99.6.323-338 Schreiber JB, Nora A, Stage FK, Barlow EA, King J (2006) Reporting structural equation modeling and confirmatory factor analysis results: A review. J Educ Res 99(6):323–338

17. Nunnally JCBI Psychometric theory 1994 New York McGraw-Hill Nunnally JCBI (1994) Psychometric theory. McGraw-Hill, New York

18. Fayers PHR Assessing quality of life in clinical trials: methods and practice 2005 New York Oxford University Press Fayers PHR (2005) Assessing quality of life in clinical trials: Methods and practice. Oxford University Press, New York

19. Brown TA (2015) Confirmatory Factor Analysis for Applied Research Second Edition. Guilford Publications, New York

20. Muthén BO Bollen KA Long JS Goodness of fit with categorical and other non-normal variables Testing structural equation models 1993 Newbury Park, CA: Sage A:Sage 205 243 Muthén BO, Bollen KA, Long JS (1993) Goodness of fit with categorical and other non-normal variables. Testing structural equation models. A:Sage, Newbury Park, CA: Sage pp 205–243

21. Cronbach LJ Warrington WG Time-limit tests: estimating their reliability and degree of speeding Psychometrika 1951 16 2 167 188 10.1007/BF02289113 14844557 Cronbach LJ, Warrington WG (1951) Time-limit tests: Estimating their reliability and degree of speeding. Psychometrika 16(2):167–188 14844557 10.1007/BF02289113

22. Cronbach LJ Meehl PE Construct validity in psychological tests Psychol Bull 1955 52 4 281 302 10.1037/h0040957 13245896 Cronbach LJ, Meehl PE (1955) Construct validity in psychological tests. Psychol Bull 52(4):281–302 13245896 10.1037/h0040957

23. Goswami P (2019) The conceptualization, development and validation of a treatment-related patient-reported outcome measure in haematological malignancies for use in clinical practice. Thesis, University of Hertfordshire

24. Chen W-H Thissen D Local dependence indexes for item pairs using item response theory JEBS 1997 22 3 265 289 Chen W-H, Thissen D (1997) Local dependence indexes for item pairs using item response theory. JEBS 22(3):265–289

25. Tavakol M Dennick R Making sense of Cronbach’s alpha Int J Med Educ 2011 2 53 55 10.5116/ijme.4dfb.8dfd 28029643 PMC4205511 Tavakol M, Dennick R (2011) Making sense of Cronbach’s alpha. Int J Med Educ 2:53–55 28029643 10.5116/ijme.4dfb.8dfd PMC4205511

26. Sztankay M, Giesinger J, Holzner B (2011) Gender differences in quality of life in patients with haematological malignancies. MEMO: Magazine Eur Med Oncol. 224–226

27. Bergerot CD Clark KL Nonino A Waliany S Buso MM Loscalzo M Course of distress, anxiety, and depression in hematological cancer patients: association between gender and grade of neoplasm Palliat Support Care 2015 13 2 115 123 10.1017/S1478951513000849 24182842 Bergerot CD, Clark KL, Nonino A, Waliany S, Buso MM, Loscalzo M (2015) Course of distress, anxiety, and depression in hematological cancer patients: Association between gender and grade of neoplasm. Palliat Support Care 13(2):115–123 24182842 10.1017/S1478951513000849

28. Paller CJ Campbell CM Edwards RR Dobs AS Sex-based differences in pain perception and treatment Pain Med 2009 10 2 289 299 10.1111/j.1526-4637.2008.00558.x 19207233 PMC2745644 Paller CJ, Campbell CM, Edwards RR, Dobs AS (2009) Sex-based differences in pain perception and treatment. Pain Med 10(2):289–299 19207233 10.1111/j.1526-4637.2008.00558.x PMC2745644

29. Kessler RC Berglund P Demler O Jin R Merikangas KR Walters EE Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National comorbidity survey replication Arch Gen Psychiatry 2005 62 6 593 602 10.1001/archpsyc.62.6.593 15939837 Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE (2005) Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National comorbidity survey replication. Arch Gen Psychiatry 62(6):593–602 15939837 10.1001/archpsyc.62.6.593

30. Pud D Gender differences in predicting quality of life in cancer patients with pain Eur J Oncol Nurs 2011 15 5 486 491 10.1016/j.ejon.2010.12.005 21256079 Pud D (2011) Gender differences in predicting quality of life in cancer patients with pain. Eur J Oncol Nurs 15(5):486–491 21256079 10.1016/j.ejon.2010.12.005

31. Immanuel A Hunt J McCarthy H van Teijlingen E Sheppard ZA Quality of life in survivors of adult haematological malignancy Eur J Cancer Care (Engl) 2019 28 4 e13067 10.1111/ecc.13067 31020751 Immanuel A, Hunt J, McCarthy H, van Teijlingen E, Sheppard ZA (2019) Quality of life in survivors of adult haematological malignancy. Eur J Cancer Care (Engl) 28(4):e13067 31020751 10.1111/ecc.13067

32. Priscilla D Hamidin A Azhar MZ Noorjan KO Salmiah MS Bahariah K Quality of life among patients with hematological cancer in a Malaysian hospital Med J Malaysia 2011 66 2 117 120 22106690 Priscilla D, Hamidin A, Azhar MZ, Noorjan KO, Salmiah MS, Bahariah K (2011) Quality of life among patients with hematological cancer in a Malaysian hospital. Med J Malaysia 66(2):117–120 22106690

33. Slovacek L Slovackova B Pavlik V Jebavy L Health-related quality of life in acute myeloid leukaemia and multiple myeloma survivors undergoing autologous progenitor stem cell transplantation: A retrospective analysis Rep Prac Oncol Radiother 2007 12 231 238 10.1016/S1507-1367(10)60062-4 Slovacek L, Slovackova B, Pavlik V, Jebavy L (2007) Health-related quality of life in acute myeloid leukaemia and multiple myeloma survivors undergoing autologous progenitor stem cell transplantation: A retrospective analysis. Rep Prac Oncol Radiother 12:231–238

34. Goswami P Oliva EN Ionova T Else R Kell J Fielding AK Paper and electronic versions of HM-PRO, a novel patient-reported outcome measure for hematology: an equivalence study J Comp Eff Res 2019 8 7 523 533 10.2217/cer-2018-0108 31037971 Goswami P, Oliva EN, Ionova T, Else R, Kell J, Fielding AK et al (2019) Paper and electronic versions of HM-PRO, a novel patient-reported outcome measure for hematology: An equivalence study. J Comp Eff Res 8(7):523–533 31037971 10.2217/cer-2018-0108

35. Basch E Deal AM Kris MG Scher HI Hudis CA Sabbatini P Symptom monitoring with Patient-Reported outcomes during routine cancer treatment: A randomized controlled trial J Clin Oncol 2016 34 6 557 565 10.1200/JCO.2015.63.0830 26644527 PMC4872028 Basch E, Deal AM, Kris MG, Scher HI, Hudis CA, Sabbatini P et al (2016) Symptom monitoring with patient-reported outcomes during routine cancer treatment: A randomized controlled trial. J Clin Oncol 34(6):557–565 26644527 10.1200/JCO.2015.63.0830 PMC4872028

36. Efficace F Cottone F Abel G Niscola P Gaidano G Bonnetain F Patient-reported outcomes enhance the survival prediction of traditional disease risk classifications: an international study in patients with myelodysplastic syndromes Cancer 2018 124 6 1251 1259 10.1002/cncr.31193 29231969 Efficace F, Cottone F, Abel G, Niscola P, Gaidano G, Bonnetain F et al (2018) Patient-reported outcomes enhance the survival prediction of traditional disease risk classifications: An international study in patients with myelodysplastic syndromes. Cancer 124(6):1251–1259 29231969 10.1002/cncr.31193