Patient preferences in the treatment of hemophilia A: A latent class analysis.

OBJECTIVE: To examine subgroup-specific treatment preferences and characteristics of patients with hemophilia A.
METHODS: Best-Worst Scaling (BWS) Case 3 (four attributes: application type; bleeding frequencies/year; inhibitor development risk; thromboembolic events of hemophilia A treatment risk) conducted via online survey. Respondents chose the best and the worst option of three treatment alternatives. Data were analyzed via latent class model (LCM), allowing capture of heterogeneity in the sample. Respondents were grouped into a predefined number of classes with distinct preferences.
RESULTS: The final dataset contained 57 respondents. LCM analysis segmented the sample into two classes with heterogeneous preferences. Preferences within each were homogeneous. For class 1, the most decisive factor was bleeding frequency/year. Respondents seemed to focus mainly on this in their choice decisions. With some distance, inhibitor development was the second most important. The remaining attributes were of far less importance for respondents in this class. Respondents in class 2 based their choice decisions primarily on inhibitor development, also followed, by some distance, the second most important attribute bleeding frequency/year. There was statistical significance (P < 0.05) between the number of annual bleedings and the probability of class membership.
CONCLUSIONS: The LCM analysis addresses heterogeneity in respondents' choice decisions, which helps to tailor treatment alternatives to individual needs. Study results support clinical and allocative decision-making and improve the quality of interpretation of clinical data.

Introduction

Hemophilia, an inherited, rare bleeding disorder, is complex to diagnose and manage [1]. Hemophilia A is more common than B; ~8/10 people with hemophilia have type A [2]. Patients suffer from repeated hemorrhagic episodes in joints and soft tissues [3]. Patients bleed longer than other people, and bleedings can occur internally in joints and muscles, as well as externally via minor cuts, dental procedures, or trauma [4].

An increasing number of people with bleeding disorders have been identified since 1999. The World Federation of Hemophilia identified 111,203 people in 1999 with inherited bleeding disorders such as hemophilia and 337,641 in 2018 [5]. The number of hemophilia cases was 78,629 in 1999 and 210,454 in 2018. Here, the number of 173,711 hemophilia A cases accounted for about 83% of the total hemophilia cases in 2018. The proportion of hemophilia A in the total number of hemophilia cases was 53,864 in 1999 and 173,711 in 2018, which corresponds to an increase of over 200%. In Germany, medical data of patients with bleeding disorders are consolidated in the German Hemophilia Registry (Deutsches Hämophilieregister, DHR). In 2018, 4,240 people with hemophilia A and 785 people with hemophilia B were included in the registry and 2,583 (~61%) of hemophilia A and 403 (~51%) of hemophilia B cases had severe hemophilia [6].

In a first paper we analyzed patients’ preferences of a whole sample set regarding general relative importance of all attributes with a mixed logit model [7]. Here, we aimed to assess heterogeneity of patients’ preferences for alternative hemophilia A treatments in Germany; the main focus being to analyze possible differences in preference patterns in the sample regarding treatment characteristics. A Best–Worst Scaling (BWS) Case 3 was conducted in an interviewer-administered survey. Two classes of respondents with heterogeneous preferences were identified in the current latent class model analysis.

Methods

Ethics statement

Participants were recruited by an external market research company. Prior to participation, patients gave informed consent. Participation was voluntary and anonymous [7]. All documents used in the study were primarily reviewed by the ethics committee at the State Medical Association of Baden-Württemberg. The study was declared harmless and approved (F-2017-048). In addition, the study was notified to the State Medical Associations of Hessen, Lower Saxony and Saarland, as these are the chambers responsible for the study centers where patient recruitment took place. This study used an anonymous online data collection. All patients gave a virtual declaration of consent before the actual survey by clicking a correspoding box and thus agreed to participate in the research project. All participants were comprehensively informed and enlightened about the research project. The patients were free to answer the questions addressed to them or to terminate the questionnaire at any time during the survey. No IP addresses of the computers, tablets etc. used or other security relevant information was collected.

Best–worst scaling case 3

Discrete choice experiments (DCE) are quantitative methods for measuring stated preferences. They are widely accepted in healthcare, and increasingly discussed by regulatory bodies [8-11]. DCEs assume a health product/service can be described in terms of its attributes and levels of its attributes. The attributes’ relative importances are determined by analyzing tradeoffs between decision-relevant attributes and their levels. DCE participants choose between ≥2 hypothetical alternatives characterized by different attributes in repeated choice scenarios. Assuming rational choice decisions, the treatment alternative benefit and its characteristics can be analytically determined [12].

A BWS is a special form of classical DCE. In a survey using a BWS Case 3 (multi-profile case) approach there are ≥3 alternatives in a choice scenario, and respondents are required to identify the most and least preferred alternative in each. Assuming there are only three alternatives in a choice task, a full ranking of preferred alternatives can be obtained. In case of >3 alternatives, a follow-up question regarding the remaining alternatives between the previously chosen best and worst alternative is answered. The BWS is approved as a valuable way to analyze patients’ preferences [13-16].

Attributes and levels

To identify treatment attributes for the BWS survey, a literature search and qualitative pre-test interviews with 12 patients were conducted [7]. Based on these, four attributes with various levels were chosen: type of application (intravenous/subcutaneous application: 1/2/3x per week); development of inhibitors (no [0%]/ low [2%]/ medium [4%]); bleeding frequency/year (0/5/15/25 per year); risk of thromboembolic events (no [0%]/ low [1%]/ medium [2%] risk [7]).

Experimental design

The combination of attribute levels resulted in an experimental design (32 x 41 x 61; two attributes with three levels, one attribute with four, and one attribute with six) with 216 possible choice alternatives [7]. A d-efficient fractional factorial design with 240 choice scenarios was created with Ngene version 1.2.1 (ChoiceMetrics, Sydney, NSW) [17]. A dominance test was used to assess validity (rationality) of respondents’ choice decisions. Within a choice scenario, containing three treatment alternatives, respondents were asked to choose the most and least preferable alternative. Attributes were randomized to control for order effects across respondents. S1 Fig shows a screenshot of the example choice set.

Recruitment

The study has been conducted in two phases. Prior to the main survey, a qualitative preliminary study was conducted using qualitative interviews with affected patients (N = 12). After evaluation of these interviews, the main survey was conducted from October 2018 to May 2019. The participants were recruited in cooperation with an independent market research company, specialized in the recruitment of target groups in the healthcare sector. Patients with hemophilia A patients who were at least 18 years old, had adequate German language skills, and agreed to participate in the 30- to 60-minute study interview were eligible for the survey [7]. The patient survey was conducted by interview and with the aid of a computer-assisted questionnaire. The survey was accessed via a provided link. 94.7% of respondents were male, average age was 34.68 years (standard deviation: 13.5). For more information on the sample characteristics please see S1 Table.

Statistical analysis

A latent class model (LCM) was used to analyze the data [18, 19]. LCMs assume that respondents differ in terms of their preferences and that individual choice behavior depends on observable attributes and on latent heterogeneity that varies with unobservable factors. Individuals are implicitly sorted into a set of different latent classes with homogeneous preferences. The classes are latent, since it is not determined a priori which respondent belongs to which class. LCMs are semi-parametric models, which free the analyst from unwarranted distributional assumptions. The models have a discrete distribution of coefficients and are used to uncover possible different preference pattern among classes. Class membership is represented in terms of class probability, which may depend on socio-economic characteristics of the respondent, general attitudes, previous treatment experience or other factors. LCM approaches offer insights into the heterogeneity of patient preferences that are not readily identifiable through other discrete choice models, especially when there are reasons to believe that these preferences are clustered around certain values [20].

The aim of the LCM was to quantify the average impact of attribute levels on the therapy preference. For each attribute level, both the mean coefficient, standard error and the 95% confidence interval were estimated, and for each attribute the relative importance. The attribute level with the highest coefficient in each class is most decisive in the choice of a therapy. The relative attribute importance for each class was calculated using the difference between the largest and the smallest level coefficients for each attribute. A score of 10 was given to the attribute with the largest difference. All other importance scores were calculated in relation to the most important attribute. Effects coding was used for the attributes in the analysis, where the reference level is generated by the inverted sum of the remaining levels.

Analysis of the LCM was conducted with Stata 15.1 (StataCorp, College Station, TX).

Three respondents failed the rationality test using dominant choice alternatives. The analysis with and without these respondents yielded better results for the model without the failed respondents. Eventually, the data of 57 respondents were used for the final analysis.

Results

Sample population

The overall sample included 57 respondents [7]. Given the small size, two classes were used for the LCM analysis. The probability that a respondent would be assigned to a particular class depended on the sequence of respondents’ choice decisions. Average probability of class membership was 0.9884 for class 1, and 0.9698 for class 2. There was no uncertainty regarding assignment of respondents to one of the classes. Respondents were more likely to be assigned to class 1. Class 1 included 65% of the sample and class 2 included 35%. There were no statistically significant differences between classes’ sociodemographic characteristics (S2 Table for Class 1/2 and [7] for total).

LCM: Capturing heterogeneity

The two classes differed in their preferences regarding treatment of hemophilia A. Table 1 shows the latent class analysis results. The table includes the mean coefficients (utilities) along with corresponding standard errors, confidence intervals, and significance levels. Coefficients are scaled to sum to zero within each attribute, and are interpreted as relative utility of each attribute level. Higher values are associated with higher preferences. Attribute levels with positive signs were preferred over levels with negative signs by the respondents in the choice decisions. The larger the absolute value of a coefficient, the greater the impact on respondents’ choice decisions. Large negative coefficients indicate a large negative impact on choice decisions. Across classes, the signs of all coefficients were as expected. A less frequent application was preferred to a more frequent one, a smaller number of bleedings was preferred to a higher number, and a lower risk of inhibitor development and thromboembolic events was preferred to a higher risk. However, there were differences between the classes in the magnitude of the impact of the attribute and levels.

Table 1

Results of the latent class analysis.

	Class 1 (65%)					Class 2 (35%)
Level	Coef.	SE	95% CI		Sig	Coef.	SE	95% CI		Sig
Type of application
Intravenous 1x/week	0.44	0.14	0.17	0.72	***	0.28	0.19	-0.10	0.65
Intravenous 2x/week	0.03	0.13	-0.21	0.28		-0.02	0.20	-0.41	0.36
Intravenous 3x/week	-0.44	0.14	-0.72	-0.16	***	-0.84	0.21	-1.24	-0.44	***
Subcutaneous 1x/week	0.17	0.15	-0.12	0.46		1.02	0.21	0.61	1.42	***
Subcutaneous 2x/week	0.26	0.12	0.02	0.51	**	-0.17	0.20	-0.56	0.23
Subcutaneous 3x/week	-0.46	0.14	-0.74	-0.19	***	-0.26	0.19	-0.63	0.11
Development of inhibitors
No (0%)	1.13	0.11	0.92	1.35	***	2.92	0.22	2.48	3.35	***
Low (2%)	0.13	0.08	-0.03	0.30		0.46	0.14	0.18	0.73	***
Medium (4%)	-1.27	0.11	-1.48	-1.05	***	-3.37	0.27	-3.90	-2.84	***
Bleeding frequency per year
0 bleedings	2.91	0.17	2.59	3.24	***	0.89	0.19	0.52	1.26	***
5 bleedings	1.55	0.12	1.31	1.79	***	0.83	0.15	0.53	1.13	***
15 bleedings	-0.95	0.12	-1.18	-0.73	***	-0.25	0.15	-0.55	0.04	*
25 bleedings	-3.51	0.20	-3.91	-3.11	***	-1.47	0.22	-1.90	-1.03	***
Risk of thromboembolic events
No risk (0%)	0.55	0.09	0.36	0.73	***	0.81	0.14	0.55	1.08	***
Low risk (1%)	0.13	0.08	-0.03	0.30		0.15	0.11	-0.08	0.37
Medium risk (2%)	-0.68	0.09	-0.86	-0.50	***	-0.96	0.14	-1.24	-0.69	***
_cons	0.63	0.29	0.05	1.20	**

AIC = Akaike Information Criteria, BIC = Bayesian Information Criterion, CI = confidence interval, Coef. = coefficient, SE = standard error, Sig = significance.

*** P < 0.01

** P < 0.05

* P < 0.1.

ll(model): -720.766; AIC: 1491.532; BIC: 1649.525

It is noticeable that each class has one attribute with very high coefficients: “Bleeding frequency per year” for class 1 and development of inhibitors for class 2. Respondents in class 1 seemed to pay more attention to bleeding frequencies in their choice decisions. In contrast, respondents in class 2 seemed to be more focused on development of inhibitors. Compared with the other attributes within the class, these attributes have very high values for level coefficients for the first and last level. Class 1 respondents showed a strong preference for “0 bleedings” (coef. = 2.91; P < 0.01) and strongly disfavored “25” (coef. = –3.51; P < 0.01). Class 2 respondents clearly preferred “no (0%)” development of inhibitors (coef. = 2.92; P < 0.01), and disfavored “medium” (4%) development (coef. = –3.37; P < 0.01). All levels of the respective attributes were statistically highly significant. The large coefficients indicate that respondents of each class seemed to be very confident in their choice decisions regarding these attribute levels.

Another noticeable difference concerns preferences for type of application. While class 1 respondents seemed to prefer weekly intravenous applications (coef: 0.44; P < 0.01), class 2 respondents seemed to prefer weekly subcutaneous application (coef: 1.02; P < 0.01). Both classes agree that they would reject a three-times-a-week intravenous application. The objection to this application seemed to be even greater for class 2 respondents (coef: –0.84; P < 0.01).

S2 Fig shows a graphical representation of the coefficients of the LCM.

The differences in preferences become clear in Fig 1, which shows the relative importance and illustrates the focus of respondents in choice decisions (type of application was divided into “subcutaneous” and “intravenous”). Class 1 respondents have a greater preference for weekly intravenous application, whereas class 2 respondents prefer a weekly subcutaneous over intravenous application. However, the distance between weekly intravenous and subcutaneous application is smaller for respondents in class 1. These respondents are indifferent between a twice-a-week intravenous and subcutaneous application, and between a three-times-a-week intravenous and subcutaneous application. The respective coefficients are close to each other. “Development of inhibitors” had a larger impact on choice decisions of class 2 respondents. With class 2 respondents the distance from the worst level (medium [4%]) to the best level (no [0%]) is more than twice as large as class 1 respondents. It is the other way around with “Bleeding frequency per year”. Table 1 shows, that the change from the most preferred level (“0 bleedings”) to the second most (“5 bleedings”) is not decisive for the respondents in class 2. The confidence intervals for the two-level coefficients overlap. For class 1, however, the level coefficients are statistically different from each other. The gradient of the line is steeper for class 1 and flatter for class 2. This indicates that a change from one level to the other is more important for class 1 respondents. With regard to the last attribute, the two classes have similar preferences.

Fig 1

Relative attribute importance.

The importance of each attribute is derived by the range between the lowest level coefficient and the highest level coefficient. This range is then normalized on a scale of 1 to 10, where 10 is the highest value and thus the most important or preferred attribute. Class 1 respondents mainly focused on bleeding frequency, while class 2 respondents paid more attention to “Development of inhibitors” in choice decisions.

Experiences and probability of class membership

Patients were asked about the characteristics of their therapies and experiences with the disease (S1 Table). Compared with class 2, there was a larger proportion of respondents in class 1 with 0 bleedings in the last year; 27.0% of the respondents in class 1 had 0 bleedings versus only 5.0% in class 2. 43.2% of class 1 and 65.0% of class 2 indicated that maximum number of bleedings was >10. Statistical significance was identified between the severity of patients’ hemophilia and class membership, and between current state of health and class membership (P < 0.05). 78.4% of class 1 and 60.0% of class 2 suffered from severe hemophilia. Only 8.1% of class 1 suffered from moderate disease, while 40.0% of class 2 had moderate hemophilia. Mild disease only applied to class 1 (13.5%). The question about current health status (P < 0.05) was answered by 18.9% respondents of class 1 and 50.0% of class 2 with “very good”, and by 62.2% of class 1 and 20.0% of class 2 with “good”. Most respondents in class 1 (81.1%) regularly administered their current therapy. Conversely, in class 2 60.0% followed a regular drug administration schedule while 40.0% took the drugs only on demand. The majority of class 1 respondents (91.9%) had a treatment plan that provided intravenous administration. In class 2, 80.0% of the respondents used intravenous administration. Detailed cross tables are shown in S3 Table, along with cross tables to compare the latent classes and another LCM with independent variables to identify characteristics of class membership. In the LCM analysis, additional covariates were included as class-membership effects for the two classes in the model. Variables that showed a significant or an almost-significant level in the cross tables were tested as covariates in the LCM. These variables were binary and assumed to be constant across alternatives for the same respondent. Only three covariates which showed significance of at least P < 0.05, were included in the final model (S3 Table). The first covariate (cov1) was derived from the question “How many bleedings have you had in the last year?” and coded 1 if respondents had 0–2 bleedings and 0 otherwise. cov2 was derived from “What was the maximum number of bleedings?” and coded 1 if respondents stated that they suffered from >20, and 0 otherwise. cov3 was based on “How would you describe your current state of health in general?” and coded 1 if respondents answered with “very good” and 0 otherwise. Reference class is class 2. Respondents in class 1 differed significantly in terms of bleeding frequencies, maximum number of bleedings, and current state of health. Class 1 respondents had a lower number of bleedings in the last year (coef. = 2.04; P < 0.05). Respondents of class 2 were more likely to have a maximum number of bleedings of >20 (coef. = –2.56; P < 0.01). Regarding current state of health, class 1 had a significantly lower proportion of respondents with a very good health state (self-report by respondents) (coef. = –2.91; P < 0.01) than class 2.

Discussion

People often form groups or segments, also latent classes, with similar interests and needs and seek similar benefits from health providers. Health organizations need to understand whether the same health treatments, prevention programs, services, and products should be applied to everyone in the relevant population or whether different treatments need to be provided to each of several groups that are relatively homogeneous internally but heterogeneous among groups. Using panel data from discrete choice experiments, latent class analysis is commonly performed to identify subsets of participants with homogeneous preferences within groups and heterogeneous preferences between groups [21]. Classification of respondents into groups or clusters is usually based on the patterns of outcome variables such as individual choice decisions in stated-preference surveys using discrete choice experiments [22].

Hemophilia A is a severe, chronic disease that makes high demands of patients’ treatment compliance. Providing therapy is very complex and requires a balance between possible benefits and risks. It is important to know the preferences and wishes of patients with regard to the treatment alternatives and application schemes of the drugs. The practical approach provided by the BWS can help improve communication between patients and their care providers, support clinical and allocative decision-making, and improve the quality of interpretation of clinical data [7]. Therapies can also be made to be more patient-oriented based on the knowledge gained from this method. As such, care can be made more effective and efficient; increasing benefits for patients [7]. Two classes of respondents with different preferences were identified in the current LCM analysis. Overall, results showed that patients attach the highest importance to "Bleeding frequency per year" and risk of “Development of inhibitors”. A higher frequency of bleeding and a higher risk of inhibitor development would significantly impact patients’ choice decisions. Class 1 respondents made their choice decisions mainly with focus on “Bleeding frequency per year” while class 2 respondents paid more attention to “Development of inhibitors”. The “Type of application” did not seem to influence the choice decision much compared with other attributes. In this respect, however, there was a difference between the two classes. Regarding class characteristics, the two seemed to be segmented in terms of experiences with the disease. The classes could be described in terms of patients’ experience with the disease and current state of health.

Heterogeneous preference structures may influence the acceptance, and thus the adherence, of patients to alternative therapies. Socio-demographic variables, experiences and attitudes can cause heterogeneous preference structures. With latent class analysis, these preference structures can be revealed and divided into segments. The distribution of patients in both latent classes is tabulated (see S2 Table). Comparison of the sociodemographic characteristics of the patients in both latent classes by chi-square tests revealed no significant differences between class members. However, there were significant differences between class members regarding treatment experience, number of bleeds, maximum number of bleeds, type of current treatment, severity of hemophilia A, and self-reported health status. In our analysis, we found that, compared with class 1, class 2 contained a significantly larger proportion of patients with more than 2 bleeding episodes per year. In addition, there were more patients in this class who had experienced a maximum number of bleeding episodes greater than 20. For these patients in class 2, avoiding the development of inhibitors was the most decisive criterion when choosing a therapy alternative. Patients who had more bleeds per year placed greater weight on the attribute of development of inhibitors when choosing a therapy. Avoidance of inhibitors was more important for patients in class than for patients in class 1. Based on this information, physicians can communicate and target therapy.

Limitations

A limitation is the small sample size. Due to the rare nature of the disease and the low prevalence in Germany, it was difficult to recruit patients to achieve a larger size. The sample of N = 57 and therefore the sizes of the latent classes tend to be towards the lower limit. This is also reflected in the high standard errors. A larger sample size is recommended for LCM; most studies in other indications include >300 respondents [21].

There are two types of heterogeneity in discrete choice data. One type is called taste (or preference) heterogeneity, and the other is called scale heterogeneity. Taste heterogeneity identifies groups of respondents who like or dislike different alternatives in a systematic and quantifiable way. Scale heterogeneity refers to the error variance which is thought to vary systematically in response to task complexity, e.g., the number of choice alternatives. That is, heterogeneity cannot be linked to measurable aspects of the decision maker. Scale heterogeneity can be an issue, particularly when comparing coefficients from datasets from different populations or data generated from different sources, e.g., sources using different sampling strategies [23-25]. As we did not intend to analyze how patients made a decision, but rather what kind of decision they made, in our study we did not separate preference heterogeneity from scale heterogeneity. The main interest of the study was to investigate systematic preference heterogeneities, e.g., whether there are differences in therapy uptake in the group of hemophilia patients and which therapy characteristics are accounting for the difference. Since we examine a very small and homogeneous patient population in the study, the analysis of scale heterogeneity was omitted.

Conclusions

This study analyzed patients’ preferences for hemophilia A treatments using a BWS. Data were analyzed using the LCM approach, which addresses heterogeneity in respondents’ choice decisions. Study results identified two classes of respondents with different preferences: respondents of class 1 made their choice decisions mainly with a focus on the attribute “Bleeding frequency per year,” while respondents of class 2 paid more attention to “Development of inhibitors.” This helps in decision-making to tailor treatment alternatives for hemophilia A patients to individual needs.

Screenshot of the example choice set.

The question is: “You are diagnosed with hemophilia A. The doctor asks you to choose between therapy A, therapy B and therapy C. In your opinion, which therapy is the best and which is the worst?”.

(DOCX)

Click here for additional data file.

Coefficients of the latent class model.

Coefficients are displayed separately for each of the two classes. Vertical bars around the coefficients represent the 95% confidence interval. When the confidence intervals overlap for adjacent levels within an attribute, the coefficients of these levels are statistically not different from each other. The upper graphic shows coefficients of class 1 respondents, the lower graphic coefficients of class 2 respondents. The first attribute shows the two types of application, intravenous and subcutaneous, on top of each other. Also, the greater distance between the values “5 bleedings” and “15 bleedings” as well as between “15 bleedings” and “25 bleedings” compared with the distance between 0 and 5 bleedings of the attribute “Bleeding frequency per year” is taken into account in the graph.

(DOCX)

Click here for additional data file.

Characteristics of the therapy and experiences.

* P < 0.1, ** P < 0.05, *** P < 0.01.

(DOCX)

Click here for additional data file.

Sociodemographic characteristics of the overall sample and the two latent classes.

(DOCX)

Click here for additional data file.

Latent class model with membership variables.

CI = confidence interval, Coef. = coefficient, LCM = latent class model, SE = standard error, Sig = significance. *** P < 0.01, ** P < 0.05, * P < 0.1. ll(model): -709.411; AIC: 1474.821; BIC: 1651.773; degrees of freedom: 28. In the LCM analysis additional covariates were included as class-membership effects for the two classes in the model. Variables that showed a significant or an almost-significant level in the cross tables (see S1 Table) were tested as covariates in the latent class model. These variables were binary and assumed to be constant across alternatives for the same respondent. Only three covariates which showed a significance of at least P < 0.05 were included in the final model. The first covariate (cov1) was derived from the question “How many bleedings have you had in the last year?” and coded 1 if respondents had 0 to 2 bleedings last year and 0 otherwise. Second covariate (cov2) was derived from the question “What was the maximum number of bleedings?” The variable was coded 1 if respondents stated that they suffered from a maximum number of bleedings of more than 20 and 0 otherwise in the dataset. The third covariate (cov3) based on the question “How would you describe your current state of health in general?” It was coded 1 if respondents answered with very good and coded with 0 otherwise. Reference class is class 2. Respondents in class 1 differed significantly in terms of bleeding frequencies, maximum number of bleedings, and current state of health. Class 1 respondents had a lower number of bleedings in the last year (coef. = 2.04; P < 0.05). Respondents of class 2 were more likely to have a maximum number of bleedings of more than 20 (coef. = -2.56; P < 0.01). Regarding current state of health, class 1 had a significant lower proportion of respondents with a very good health state (self-report by respondents) (coef. = -2.91; P < 0.01) than class 2 respondents.

(DOCX)

Click here for additional data file.

9 Apr 2021

PONE-D-20-34726

Patient preferences in the treatment of hemophilia A: A Latent Class analysis

PLOS ONE

Dear Dr. Mühlbacher,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please note that while you will not see a long list of reviewer comments, a major revisions decision was appropriate given that the work involved in addressing them appeared to be a heavy lift. Each of the comments appears to be reasonable from my perspective, and if addressed appropriately, would improve the paper. Please be sure to address each carefully.

Please submit your revised manuscript by May 24 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Damian Adams

Academic Editor

PLOS ONE

Additional Editor Comments:

The paper has received two reviews, both encouraging revision and resubmission. Based on their comments, I believe a major revisions decision is appropriate, and would like to see the authors resubmit after carefully addressing each of the reviewers' comments and concerns, which appear reasonable. Upon resubmitting, please be sure to provide a detailed indication of how each comment and concern was addressed, either indicating changes made to the paper or otherwise explaining your approach. Since this is a major revisions decision, please note that the resubmitted manuscript will likely need to be assessed by the reviewers again. Thank you.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for providing the following Funding Statement:

"This study was financed by Roche Pharma AG (https://www.roche.de). Employees of the sponsor are listed as authors and were involved in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

All authors received support for third-party editing assistance, provided by Roche Pharma AG, Grenzach-Wyhlen, Germany. Mr. Sadler (M.Sc.) reports grants from Roche Pharma AG, Grenzach-Wyhlen, Germany, during the conduct of the study. Prof. Dr. Mühlbacher reports grants from Roche Pharma AG, Grenzach-Wyhlen, Germany, during the conduct of the study. Dr. Lamprecht is an employee of Roche Pharma AG, Grenzach-Wyhlen, Germany. Ms. Juhnke reports grants from Roche Pharma AG, Grenzach-Wyhlen, Germany, during the conduct of the study."

We note that one or more of the authors is affiliated with the funding organization, indicating the funder may have had some role in the design, data collection, analysis or preparation of your manuscript for publication; in other words, the funder played an indirect role through the participation of the co-authors.

If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study in the Author Contributions section of the online submission form. Please make any necessary amendments directly within this section of the online submission form.  Please also update your Funding Statement to include the following statement: “The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If the funding organization did have an additional role, please state and explain that role within your Funding Statement.

Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

Within your Competing Interests Statement, please confirm that this commercial affiliation does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If this adherence statement is not accurate and  there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

3. Thank you for stating the following in the Competing Interests section:

"All authors received support for third-party editing assistance, provided by Roche Pharma AG, Grenzach-Wyhlen, Germany. Mr. Sadler (M.Sc.) reports grants from Roche Pharma AG, Grenzach-Wyhlen, Germany, during the conduct of the study. Prof. Dr. Mühlbacher reports grants from Roche Pharma AG, Grenzach-Wyhlen, Germany, during the conduct of the study. Dr. Lamprecht is an employee of Roche Pharma AG, Grenzach-Wyhlen, Germany. Ms. Juhnke reports grants from Roche Pharma AG, Grenzach-Wyhlen, Germany, during the conduct of the study."

We note that one or more of the authors are employed by a commercial company: Gesellschaft für empirische Beratung GmbH.

3.1. Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form.

Please also include the following statement within your amended Funding Statement.

“The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

3.2. Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

Within your Competing Interests Statement, please confirm that this commercial affiliation does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests) . If this adherence statement is not accurate and  there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

4. We noted in your submission details that a portion of your manuscript may have been presented or published elsewhere.

"Overall experimental design (Muhlbacher et al, Value Health, 2020).

This does not constitute dual publication as in the first paper we analyzed patients’ preferences of a whole sample set regarding general relative importance of all attributes with a mixed logit model (Muhlbacher et al, Value in Health, 2020; see uploaded paper as related manuscript file). Here, we aimed to assess heterogeneity of patients' preferences for alternative hemophilia A treatments in Germany; the main focus being to analyze possible differences in preference patterns in the sample regarding treatment characteristics."

Please clarify whether this [conference proceeding or publication] was peer-reviewed and formally published. If this work was previously peer-reviewed and published, in the cover letter please provide the reason that this work does not constitute dual publication and should be included in the current manuscript.

5. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This brief paper describes a latent class analysis on BWS case-3 responses regarding homophilia A treatment preferences of 57 respondents. Unlike their primary analysis, this paper describes two classes. This paper represents a modest contribution based on the topic and interpretation. It does perform above the basics or introduce innovations in methods, practice, or application. As an exploratory analysis,

Over half of this secondary paper re-iterates description found in the primary one. Its primary contribution is the application of the stata LCM package to examine heterogeneity. The authors provide a modest interpretation of the grade of membership and differences in attribute importance. They did not include any qualitative evidence to bolster the explanation and did not conduct a confirmatory analysis.

Major comments:

1. Please describe the respondents in detail, including aspects of their recruitment, characteristics, and interview experience, which may affect their grade-of-membership and the implications of the results.

2. Please include screenshots of the full survey instrument. These screenshots are typically mandatory prior to the review of any stated preference evidence.

3. Please emphasize the difference between taste and scale in the interpretation of LCM results.

Reviewer #2: The author conducts a latent class analysis for identifying hidden classes for patient preferences for hemophilia A. The research question is an interesting and important one. But i have some methodological queries. The author has failed to mention how their results will be used to identify hidden cohorts in the real world. When we conduct a latent class analysis, we identify hidden classes. It is to be understood by varying the variables involved in latent class model formation; we will identify different classes. So after having the rationality for the selection of questions, a latent model needs to be formed. (until here, the author has conducted though rationality for final question utilized can be elaborated further from a clinical and statistical standpoint). After the hidden population has been identified, they need to extrapolate the identified class variable (either 0 or 1) to the primary data frame and make a univariate analysis (on the unused variables) to distinguish between the identified population. The author also mentions that there seems to be no difference in demographics between the latent class. How will we use it in the real world to differentiate between the hidden class so that the healthcare worker can plan for interventions to address the issues? I want the author to address this critical question.

The author need to elaborate the discussion. The discussion is too short. Elaborate on past studies and papers which have adopted similar methodology in other diseases.

In the introduction, the author has written prevalence in the form of mean and sd.

Is the first paper, which the author referring to also utilizing the same data. The authors should make sure that there is no result repetition.

By means of not explaining the variable distribution to the study population, it would be difficult for readers to understand.

The writeup is more statistically oriented concentrating on non-important attributes. I would ask the author to have a thorough write-up overhaul having in mind that PlosOne has a broad readership. Also try to explain the statistical concepts then and there whereever you are mentioning.Explaining a tough concept lucidly to the reader is also an important art.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Praveen Kumar M

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

14 Jun 2021

Dear Dr. Adams,

Dear Reviewers,

On behalf of my co-authors, thank you for providing us with the opportunity to revise our manuscript for PLOS ONE. As requested, please find a point-by-point response to each of the reviewers’ queries and suggestions in the regarding uploaded document, as well as the items mentioned in the compliance form.

Thank you once again for your time; I hope to hear from you in due course.

Kind regards,

Axel C. Mühlbacher

Submitted filename: Pat_Preferences_Hemophilia_A_LClass_journal_response_210604.docx

Click here for additional data file.

10 Aug 2021

Patient preferences in the treatment of hemophilia A: A Latent Class analysis

PONE-D-20-34726R1

Dear Dr. Mühlbacher,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Damian Adams

Academic Editor

PLOS ONE

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

Reviewer #2: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!Congrats!

Reviewer #2: The author team has addressed all the queries raised to a satisfactory extent. Congrats to the team for coming up with the revision. Thanks.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Praveen Kumar M

13 Aug 2021

PONE-D-20-34726R1

Patient preferences in the treatment of hemophilia A: A Latent Class analysis

Dear Dr. Mühlbacher:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Damian Adams

Academic Editor

PLOS ONE