Adaptation and validation of the French version of the Hypoparathyroid Patient Questionnaire 28 (HPQ28) in the ComPaRe-Epi-Hypo e-cohort

Abstract

Chronic hypoparathyroidism is a rare disease associated with an impaired quality of life. Recommendations suggest frequent monitoring of quality of life, but for French-speaking people, only generic scales are available despite the fact that chronic hypoparathyroidism has specific symptoms and impact. The aim of this study was to adapt and validate the French version of Hypoparathyroid Patient Questionnaire 28 (HPQ28), an already validated tool in patients living with chronic hypoparathyroidism, available in English and German. HPQ28 was translated and back-translated from English into French. Translations were harmonized with the original author. Assessment of psychometric properties of the French version of HPQ28 was performed in the ComPaRe-Epi-Hypo e-cohort, a nationwide cohort of adult patients living with chronic hypoparathyroidism in France. Internal consistency was evaluated using Cronbach’s alpha. Dimensional validity was studied using confirmatory factor analysis (CFA). Construct validity compared the answers from the French version of HPQ28 with those from the EQ-5D-5L, EQ-5D-VAS, and MYMPO2 instruments. Reliability was evaluated by the intra-class correlation coefficient (ICC) of a test–retest within a 2-wk interval. Between August 2023 and August 2024, 183 patients completed HPQ28, EQ-5D, and MYMOP2 scales. The majority (92%) of the participants were women, with a median[IQR] age of 52[44;60]. Etiology of the disease was neck surgery and genetic abnormalities in 82% and 8% of cases, respectively. Internal consistency was good (Cronbach’s alpha 0.93, 95% CI 0.91 to 0.94). CFA found a unidimensional structure of the questionnaire. Construct validity showed positive correlation with MYMOP2 (r = 0.64) and negative correlations with EQ-5D VAS (r = −0.49) and EQ-5D-5L (r = −0.64) scores, as hypothesized. Reliability was adequate, with an ICC of 0.88 (95% CI 0.84 to 0.91). In conclusion, we adapted and validated HPQ28 for French-speaking patients suffering from chronic hypoparathyroidism. It can therefore now be used for both research and clinical follow-up.

Graphical Abstract

Graphical Abstract.

Introduction

Chronic hypoparathyroidism is a rare disorder characterized by a low blood concentration of calcium along with undetectable or abnormally “normal” levels of circulating parathyroid hormone.¹ Etiologies of chronic hypoparathyroidism are dominated by neck surgery where parathyroid glands are removed or injured.² In most cases, hypoparathyroidism is transient and resolves within a few weeks but it can also turn into a chronic form (generally after 6 mo), in which case it usually becomes permanent. Chronic complications of hypoparathyroidism include renal calcifications (nephrolithiasis and/or nephrocalcinosis) as well as calcium deposition in any tissue.³ Patients facing chronic hypoparathyroidism have a decreased quality of life (QoL) as compared with the general population,⁴ to their previous life,⁵ or to control patients who underwent the same surgery but did not develop chronic hypoparathyroidism.⁶ The latest international guidelines strongly recommend following QoL in patients suffering from chronic hypoparathyroidism.⁷ Most studies use generic tools for measuring QoL in patients with hypoparathyroidism (such as SF-36^4,8 or EQ-5D⁹). However, these tools may be too generic and may lack sensitivity to account for the specific symptoms and impact of hypoparathyroidism in patients’ lives.

Recently, 2 groups have developed hypoparathyroidism-specific patient-reported outcomes questionnaires, namely the HPES¹⁰ and the Hypoparathyroid Patient Questionnaire (HPQ),¹¹ then shortened to HPQ28.¹² The HPQ28 assesses important dimensions for patients with hypoparathyroidism (namely, depression and anxiety, loss of vitality, pain and cramps, gastrointestinal symptoms, and neurovegetative symptoms) with 28 items (each scored from 0 to 3, then summed and weighted to 100%). It therefore scores from 0 to 100 with higher scores indicating poorer disease control. The HPQ28 has been validated in its original form in German¹¹ and in English.¹² So far, no hypoparathyroidism-specific tool has been validated for a French-speaking population. Due to cultural and linguistic specificities,¹³ it is important to adapt such a tool in this population.

The aim of this study was to adapt and validate the HPQ28 instrument into French, expending its applicability to the French-speaking population of patients living with chronic hypoparathyroidism.

Materials and methods

Adaptation of the HPQ28 in French

We followed the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN)¹⁴ and report such methods following the COSMIN study design checklist¹⁵ where appropriate.

Two investigators fluent in both French and English (J.P.B. and N.G.) independently translated the English version of the HPQ28 into French. They then met to reach a consensus. Two other independent people back-translated the French version into English and met to reach a consensus. The back translated English version was compared with the initial one during a meeting involving the HPQ28 inventor (HS) to make sure translators did not add or remove any information/meaning into items. The French version of the questionnaire was implemented into the digital platform of ComPaRe (COSMIN checklist item R1). Five patients living with chronic hypoparathyroidism were invited to test the questionnaire and ensure there was no missense in the wording (COSMIN checklist items CV3 and CV7).

Assessment of the psychometric properties of the French version of HPQ28

Patients living with chronic hypoparathyroidism were recruited from the ComPaRe-Epi-Hypo e-cohort, a nationwide cohort of patients living with chronic hypoparathyroidism in France nested within the ComPaRe project¹⁶ (COSMIN checklist item GM2). Inclusion criteria (COSMIN checklist item GM3) were (1) being 18 yr old or older, (2) self-reporting chronic hypoparathyroidism, and (3) accepting to participate. All eligible patients from the ComPaRe e-cohort were invited to participate. Calls for participation were also issued by the patient association Hypoparathyroidism France, the rare disease network OSCAR, and the research network Epi-Hypo.¹⁷

Participants who agreed to participate connected on the secured platform from ComPaRe and completed the EQ-5D (both 5L and VAS),⁹ MYMOP2,¹⁸ and the French version of HPQ28¹² (COSMIN checklist items R1 and ME1). Patients did not receive any financial compensation for their participation in this study.

We calculated the HPQ28 score following the current scoring system: for items 1 to 22, each item answer was rated from 0 (not at all) to 3 (severely/nearly every day), while for items 23 to 28, each item answer was rated from 3 (not at all) to 0 (strongly). All participants had to complete every item. The sum of item scores ranged from 0 to 84 and was weighted into a 0-to-100 scoring (COSMIN checklist item R2).

Dimensional validity of the French version of HPQ28 was assessed using a confirmatory factor analysis (CFA, COSMIN checklist item SV1). Internal consistency was assessed by the Cronbach’s alpha (COSMIN checklist items IC1 and IC2). Construct validity was assessed by verifying 2 hypotheses: a positive correlation with the MYMOP2 instrument and a negative correlation with both EQ5D-5L and EQ5D-VAS (COSMIN checklist items CriV2, ConV1, ConV3, ConV4, Resp1, Resp3, and resp4). Reliability was assessed by a test–retest with a 2-wk interval (this time interval was chosen for symptoms to remain the same, COSMIN checklist items CV5 and Resp5) and represented by using the intraclass correlation coefficient (ICC; COSMIN checklist item R2). As normal distribution could not be assumed for all variables, continuous data are presented as median[IQR] and categorical variables as effectives(percentages). Nonparametric tests were used. A p-value <0.05 indicated statistical significance (COSMIN checklist item GM6).

As participants were required to answer HPQ28 but not all other questionnaires, some data may be missing: the total numbers of available data are reported within the manuscript (COSMIN checklist item GR1). All analyses were performed using R Statistical Software (v4.2.2; R Core Team 2021).

Results

Adaptation of HPQ28 in French

The inventor of the HPQ28 tool (HS) was involved in the adaptation process. Translations from English to French were consistent with few discrepancies. Back translations were also very consistent with a few items that had to be refined. The translated tool was tested by 5 patients: none had difficulties to answer it and no misunderstanding was observed (COSMIN checklist items CV3 and CV7). All authors approved the final version of the French version of the HPQ28 questionnaire.

Assessment of the psychometric properties of the French version of HPQ28

As of 31 August 2024, the ComPaRe-Epi-Hypo cohort counted 224 patients (Figure 1), of which 220 accepted to participate in this study. Finally, 183 (83%) participants completed the French version of the HPQ28 questionnaire. Participants were mainly women with a median[IQR] age of 52[44;60] (Table 1). Most of participants (164, 90%) had at least a high school diploma and the majority (104, 57%) were employed. The majority (147, 82%) of participants had a chronic hypoparathyroidism due to neck surgery with a median[IQR] duration of 10[5;16] yr, with a minimum duration of 1 yr and a maximum duration of 47 yr. Treatment involved mainly calcium salts (mostly, calcium carbonate) and hydroxylated forms of vitamin D (mainly, alfacalcidol) and 35 (19%) participants were treated with teriparatide injections. Altogether, two-thirds (129, 65%) of participants had to take 3 or more drugs every day to treat their chronic hypoparathyroidism (Table 1). Over one-third (59, 33%) of participants had to travel more than 50 km (ie, >100 km round trip), up to 1202 km to visit the physician who took care of their chronic hypoparathyroidism, being mainly endocrinologists or nephrologists.

Figure 1.

Flow chart of the study. The ComPaRe hypoparathyroidism (HypoPT) cohort consisted in 224 people willing to participate. Out of them, 37 did not complete the first questionnaire and 4 refused to participate. Finally, 183 complete the first instance of the Hypoparathyroid Patient Questionnaire 28 at day 0 (d0) and 139 of them completed it a second time at day 15 (d15).

Table 1.

Characteristics of participants.

Characteristic	n = 183
Gender, women n (%)	168 (92)
Age (yr)
Median [IQR]	52.0 [43.5;59.5]
Range	23.0-75.0
Highest graduation, n (%)
No diploma	3 (2)
National vocational qualifications	15 (8)
High school diploma	71 (39)
Master’s degree/PhD	86 (47)
General certificate	4 (2)
Other diploma	3 (2)
(Missing)	1
Employment, n (%)
Unemployed	22 (12)
In education	2 (1)
Disabled/long-term sick	19 (10)
Employed	104 (57)
Retired or pre-retired	27 (15)
Other	8 (4)
(Missing)	1
Cause of hypoparathyroidism, n (%)
Neck surgery/radiotherapy	147 (82)
Inherited disease	14 (8)
Autoimmune/infiltrative disease	5 (3)
Unknown/missing	17 (7)
Age at disease onset (yr)
Median [IQR]	41.0 [33.0;48.5]
Range	1.0-67.0
(Missing)	4
Duration of disease (yr)
Median [IQR]	10.0 [5.0;15.5]
Range	1.0-47.0
(Missing)	4
Hypoparathyroidism-specific treatment, n (%)
Calcium salts	130 (71)
Active vitamin D (alfacalcidol or calcitriol)	146 (80)
Native vitamin D (cholecalciferol or ergocalciferol)	121 (66)
Calcium-rich diet	88 (48)
Magnesium salts	76 (42)
Diuretics	23 (13)
Teriparatide	35 (19)
Phosphate binders	1 (1)
Number of classes of treatment, n (%)
None	7 (4)
1-2	57 (31)
3-4	105 (57)
5+	14 (8)
Specialty of the main physician involved in treating hypoparathyroidism, n (%)
Endocrinology	89 (50)
Nephrology	54 (30)
Primary care	19 (11)
Other	17 (9)
(Missing)	4
Distance traveled for the consultation (km)
Median [IQR]	20.0 [7.5;100.0]
Range	0.00-1202.00
(Missing)	4
Distance traveled for the consultation, n (%)
<50 km	120 (67)
[50-100] km	17 (10)
>100 km	42 (23)
(Missing)	4

In total, the median[IQR] HPQ28 score was 47.6[36.3;61.9] points, and ranged from a minimum of 2.4 to a maximum of 94.0 (Figure 2, COSMIN checklist items GR2 and ME2). We did not find any floor or ceiling effect. Internal validity was excellent with a Cronbach’s alpha of 0.93, with a 95% CI from 0.91 to 0.94 (COSMIN checklist item IC2). Dimensional validity was assessed by the CFA (Figure 3, COSMIN checklist item CCV2).

Figure 2.

Distribution of the Hypoparathyroid Patient Questionnaire 28 (HPQ28) score among participants. The histogram of the 183 HPQ28 scores among participants at day 0 (d0) exerts almost a normal distribution.

Figure 3.

Confirmatory analysis. This graph represents the distribution of eigenvalues depending on the number of identified dimensions among the 183 Hypoparathyroid Patient Questionnaire 28 scores collected at day 0. One dimension was able to explain almost all the instrument.

Construct validity was assessed by verifying hypotheses between the HPQ28 score and patients’ QoL using the EQ-5D instrument and perception of their symptoms using the MYMOP2 instrument (COSMIN checklist item ConV1). As shown by Figure 4, we found a negative relationship between HPQ28 and the EQ5D-5L score (Figure 4A, r = −0.64, p < .001), a negative relationship with the EQ5D-VAS score (Figure 4B, r = −0.49, p < .001), and a positive relationship with the MYMOP2 score (Figure 4C, r = +0.64, p < .001), as hypothesized (COSMIN checklist items ConV3, Conv4, ConV5, Resp8, and Resp9). Altogether, a poorer QoL as assessed by nonspecific tools was associated with a poorer HPQ28 score.

Figure 4.

External validity. For assessing external validity of the Hypoparathyroid Patient Questionnaire 28 (HPQ28) measurement, we plotted EQ5D-5L (A), EQ5D-VAS (B), and MYMOP2 (C) scores against HPQ28 scores. Integrated linear regression with 95% CIs is pictured as gray areas as well as Pearson correlation factor (r) and p-value (p) for all analyses.

During retest, we obtained 139 patients (76% of the cohort, COSMIN checklist item GR3) who answered the questionnaire a second time. Median[IQR] delay between test and retest was 17[16;26] d. We found a slight but significant (p = .009) decrease of the HPQ28 score at day 15 (median[IQR] of 48[35;61] points) as compared with d0 (median[IQR] of 50[37;64] points, COSMIN checklist items Resp7 and Resp8). Results also show an excellent consistency between those 2 measures as pictured by the Blant–Altmann plot (Figure 5) as well as an ICC of 0.88 with a 95% CI from 0.84 to 0.91 (p < .001, COSMIN checklist item R2).

Figure 5.

Internal validity by Blant–Altman. The Blant–Altman graph shows a good reproducibility of Hypoparathyroid Patient Questionnaire 28 measurements 15 d apart with only a few differences between measurement outside of 2 standard derivations (pictured by the upper and lower dashed lines), whatever the average between the 2 measurements.

Discussion

A total of 343 million people speak French worldwide¹⁴; this number should increase up to 700 million by 2050. French is the everyday language of 240 million people nowadays. Of them, countless may live with chronic hypoparathyroidism without a dedicated tool to measure their QoL, so far. In this study, we adapted the HPQ28 tool in French and demonstrated its validity and reliability, underlining its suitable use for both clinical practice and research (COSMIN checklist item D5).

To do so, we followed international guidelines¹⁴: out of the checklist provided, we validated 55/71 (77%) of all items. For instance, title and abstract are fully in accordance with such recommendations but due to our study design, we did not ask participants/physicians for item relevance (CV1) or comprehensiveness (CV2), or the relevance (CV4) or comprehensiveness of results (CV6), we did not use any term that we had to provide a definition (I5) for, and we did not plan to perform neither a post-hoc analysis (GM8), an item response theory analysis (SV2 and CCV3), nor to use comparator groups (CCV1, ConV2, and Resp2). As HPQ28 is not a dichotomous (IC3), a categorical (CriV3), or a criterion (CriV1) score, some items were not relevant either. Finally, we did not have to use methods to improve its reliability (R3). Notably, we found a good but not perfect correlation of HPQ28 with 2 nonspecific Patient-Reported Outcomes questionnaires (EQ-5D and MYMOP2): on the one hand, this good correlation argues for HPQ 28 to measure QoL; on the other hand, its nonperfect correlation underlines that part of the variance of HPQ 28 is not shared with neither EQ-5D nor MYMOP2, and is therefore a specificity of the HPQ 28 tool. We could have used the SF36 like in other studies^4,8 but it would have added 36 more items to participants and be costly while we already had tools as good as this one. One could argue that we did not assess its association with biological or clinical values: here, we focused mainly on the psychometric performances of the tool; in such cases, we used other validated psychometric tools to assess its reliability. Of note, we found that patients scoring higher with HPQ28 tend (p = .07) to have more comorbidities than those with lower HPQ28 scores, highlighting its ability to measure the severity of patients condition.

We chose the HPQ28 tool because of its free access as compared with the HPES tool. The latter has also been extensively studied during chronic hypoparathyroidism^10,19,20 including a recent interventional clinical trial.²¹ HPES translation in French has not been published and would not be as easy to access as the HPQ28, which is an important point considering the deployment of Patient-Reported Outcomes in routine care.²² Moreover, HPQ28 has recently been associated with “brain fog,” a symptom reported by some patients living with chronic hypoparathyroidism, which is now known to be associated with brain changes.²³ Brain fog is characterized by “reduced concentration and reduced ability to perform day-to-day tasks”²⁴ and can be considered as a “shadow on the mind.”²⁵ It can be of great importance for patients to be considered by physicians but may be quite difficult to characterize as it is no easy to measure. Of note, during the translation process in our study, the very one concept that patients added to the questionnaire (and had to be removed) was the reference to the brain fog, highlighting its importance for patients. We would therefore recommend conducting a study in which both HPQ28 and HPES questionnaires would be used to assess their transposability and/or their uniqueness (COSMIN checklist item D6).

QoL is becoming also a very important outcome to follow as new treatments have shown to specifically alleviate the burden of chronic hypoparathyroidism.²¹ Other studies have already shown the increase of symptoms such as depression and anxiety when QoL would decrease during chronic hypoparathyroidism.²⁶ Cardiac²⁷ and mental²⁸ symptoms have already been reported to be related to QoL in other studies. In such cases, the MYMOP questionnaire is more suitable for collecting data about symptoms: one single questionnaire will probably never fit all the needs, so disposing of reliable tools for following patients living with chronic hypoparathyroidism is crucial. One team has developed the Hypoparathyroidism Symptom Diary,^29,30 which has been evaluated³¹ in clinical trials. Those tools have to be very easy to use, either for the physicians or for the patients, and therefore need to be evaluated also in a real-world setting, as even in our short experiment we faced loss-to-follow-up (ie, not 100% of participants answered twice the HPQ28), probably due to the burden of answering a questionnaire: further studies should elucidate the importance of giving a better interface for participants³² as well as addressing patient concerns regarding privacy and security³³ (COSMIN checklist item D6). One should not give up as exposing patients to Patient-Reported Outcomes may increase their willingness to respond.³⁴

Finally, our study has some strengths: namely, its fast enrolment for such a rare disease (all patients were recruited within a few months), its representativity (the study population characteristics are similar to that of most chronic hypoparathyroidism populations reported so far) as well as the fact that patient representatives participated from the very beginning of this journey (in such, they could advocate for designing and conducting this research, as well as for interpreting its results in a patients’ perspective and meaningful way), and its exhaustivity (patients completed every item of the HPQ28).

Our study also suffers from limitations. We did not calculate a priori any number of subjects to include (COSMIN checklist item GM5); therefore, our study may lack power but we used several psychometric tools to temper that limitation. Previous research has nonetheless shown that, taking into account that our psychometric scale is made of less than 30 items, 100-150 participants is enough to reach sufficient power.³⁵ We also did only a translation instead of a cultural adaptation (which would have required specific interviews³⁶) but we were able to include both the inventor of the initial tool and experts of the disease and patient representatives, both of which have helped an excellent transposition of this Patient-Reported Outcome. Finally, we considered hypoparathyroidism to be chronic if it persisted 6 mo after surgery, although 12 mo is nowadays being more consensual. We used that cutoff when we started the study but the minimum duration of hypoparathyroidism was 1 yr in this study (cf.  Table 1), therefore all the included patients had at least a 12-mo duration of hypoparathyroidism at the time they participated in this study, that is, all had chronic hypoparathyroidism as per this more stringent definition.

Altogether, this argues for a wide use of the French version of HPQ28 during chronic hypoparathyroidism for both clinical and research use. QoL is probably one of (if not) the most important outcomes to improve in chronic hypoparathyroidism as well as in any condition. Patients-physicians-researchers collaborative efforts should provide the best tools to measure and improve it.

Author contributions

Jean-Philippe Bertocchio (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing), Jessica Soyer (Data curation, Formal analysis, Writing—original draft), Natalie Grosset (Conceptualization, Investigation, Resources, Writing—original draft, Writing—review & editing), Delphine Bessonies (Conceptualization, Investigation, Resources, Writing—original draft, Writing—review & editing), Christelle Nidercorn (Conceptualization, Investigation, Resources, Writing—original draft, Writing—review & editing), Coralie Sido (Conceptualization, Investigation, Resources, Writing—original draft, Writing—review & editing), Viet-Thi Tran (Conceptualization, Methodology, Project administration, Software, Validation, Writing—review & editing), Leslie TOKO-KAMGA (Investigation, Project administration, Software, Supervision), Isabelle Pane (Data curation, Software), Akram Hecini (Data curation, Formal analysis, Writing—original draft), Heide Siggelkow (Formal analysis, Methodology, Validation, Writing—original draft, Writing—review & editing), and Pascal Houillier (Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Validation, Writing—original draft, Writing—review & editing).

Funding

This study received a financial support by the rare disease network OSCAR.

Conflicts of interest

S.J., H.A., and J.P.B. work at SKEZI, a digital platform that develops tools for collecting patient-reported outcomes. J.P.B. received research grant from Amolyt Pharma, an AstraZeneca company, and consulting fees from Takeda, Amolyt, and Ascendis Pharma.

Data availability

Data will be made available upon reasonable request to the corresponding author.