We report the first validation of a disease-specific English-language patient-reported outcome instrument for use in hypertension. The successful translation and validation of the instrument was completed in accordance with accepted standards(21, 22, 28).
Evaluation of the instrument demonstrated an important floor effect though no ceiling effect. This is in-keeping with the initial evaluation of the Spanish MINICHAL instrument(17) and reflects a relatively low symptom burden for the majority of subjects with hypertension. Internal consistency for the English MINICHAL instrument comfortably met current standards for use at group level, with dimension analysis for StM and SM finding similar values to those reported for the Spanish iteration of the instrument (0.81 and 0.75 versus 0.87 and 0.75 respectively)(17). A similar situation was observed for test-rest reliability and, in this case, the English version of the instrument also compared favorably with the Spanish instrument.
Construct validity was confirmed through the instrument’s correlation with generic instruments (EQ-5D-5L summary index and EQ-5D-5L linear scale) and the disease-specific Bulpitt-Fletcher instrument. In terms of strength of association, this was greatest with the EQ-5D-5L summary index and weakest with the Bulpitt-Fletcher instrument. This is not completely surprising considering that the Bulpitt-Fletcher instrument, though developed specifically for hypertension, is not an ideal instrument for measuring quality of life because of its mixed clinimetric-psychometric approach. A higher correlation between two instruments whose main focus is HRQoL can be therefore expected(31).
Responsiveness was tested through administration of the instrument before and after 18 weeks’ intensive treatment of hypertension, using medications and medication combinations recommended in current international guidelines, though over an accelerated timeframe(27), an intervention that seemed to have a measurable significant impact on genetic quality of life as measured through the visual analogue scale of the EQ-5D-5L (large effect size), but not on the EQ-5D-5L index (small effect size). The effect size for the StM scale was larger than for the EQ-5D-5L index, but not statistically significant due to a high dispersion of scores, and was negligible for the SM scale. It must be noted that scores were already very low for the latter scale, suggesting that the floor effect observed in this group of patients may have limited our ability to detect improvement.
A meta-analysis of cross-sectional studies has found that HRQoL is impaired across all eight domains of the SF-36 and SF-12 instruments in those with hypertension, when compared with normotensive individuals(32). Subsequent investigation has found reduced HRQoL in patients treated for hypertension, when compared with untreated hypertensive subjects(33), which may be related in large part to the subjects’ awareness of their diagnosis(34). However, these conclusions are limited by the inherent bias imparted by the cross-sectional nature of their design. In terms of longitudinal studies, which enable subjects to act as their own control group thereby minimizing confounding factors, improvement in HRQoL following treatment of hypertension has been demonstrated in a meta-analysis(30). This observation can be found with a variety of antihypertensive agents(35). Our report of improved HRQoL following treatment is therefore in-keeping with previous longitudinal data. Moreover, given that visits were predominantly delivered by allied healthcare professionals within the clinical study, our finding mirrors that of a recent Cochrane review, which concluded that HRQoL, in particular the physical functioning domain, improves with treatment of hypertension delivered by allied healthcare professionals(36). This would be also in line with the lack of responsiveness observed in this study for the SM scale.
In addition, our intervention was delivered in an accelerated timeframe, a manner of treatment delivery which is known to improve HRQoL in other fields of medicine, such as hip arthroplasty(37) and radiotherapy in the treatment of breast cancer(38). It would therefore be reasonable to propose that the improvement in HRQoL conferred through treatment of hypertension will have been accentuated by the rapid treatment protocol employed in this study.
Conversely, administration of the Bulpitt-Fletcher instrument within the protocol found a significant reduction in HRQoL following hypertension treatment. Notably, psychometric evaluation with measurement of internal consistency, floor effect, ceiling effect and construct validity for the instrument has not been reported. Additionally, test-retest reliability has only been examined for selected concepts within the questionnaire, rather than the instrument itself(39). However, responsiveness to change of the Bulpitt-Fletcher questionnaire has previously been reported through administration of the instrument within clinical studies, such as a trial comparing hypertension treatment with verapamil versus propranolol(40) and a further study comparing captopril with atenolol(41). Although generic instruments were co-administered with the Bulpitt-Fletcher instrument within these studies, no direct statistical comparison was conducted and therefore construct validity was not evaluated.
Dimensions analysis of the MINICHAL instrument results revealed a nominally greater responsiveness within the StM dimension compared to the SM dimension. Different weighting between the EQ-5D-5L and the Bulpitt-Fletcher instruments in terms of somatic symptoms versus psychological well-being may also therefore explain the discrepant results between these two instruments when applied to subjects within the study.
Study limitations
The conclusions drawn from the study are limited by the relatively low number of subjects enrolled, a drawback which could be addressed through further deployment of the translated English language MINICHAL instrument in future studies of hypertension treatment. Furthermore, as this was a before-and-after study, the effects of time, rather than treatment, on HRQoL cannot be discounted from the analysis, though this is limited by the relatively short 18-week treatment phase.
In addition, it is acknowledged that our study cohort was geographically limited to south-west England. Nevertheless, region-specific language is not used within the translated instrument and no difficulties with comprehension or cultural applicability are envisaged should the instrument be used across the United Kingdom.
The evaluation of the MINICHAL instrument alongside the EQ-5D-5L linear scale has demonstrated the latter, generic instrument to be more responsive to change than our disease-specific instrument, the converse to the expected. The reason for this is unclear, though may relate to the greater range of responses afforded by the EQ-5D-5L analogue scale in comparison with the MINICHAL’s 4 response options. Furthermore, given the brevity of the MINICHAL instrument and specific nature of the questions, some subtle features of wellbeing may not be captured with this instrument, compared to a generic assessment of HRQoL. This is particularly pertinent given the broad range of medication side effects which can result from pharmacological antihypertensive treatment, which would be difficult to capture entirely with a specific questionnaire short enough to be an acceptable burden to patients. Analysis of the components of the EQ-5D-5L summary index which changed most during treatment determined that the pain/discomfort and anxiety/depression items returned differing scores most frequently. Although anxiety/depression is covered well by the MINICHAL questions, only one question pertains to pain (“Have you experienced any pain in your chest without physical exertion?”). Therefore, a relative deficiency of the MINCHAL instrument in exploring this aspect of symptoms, together with the limited sample size in the present study, may in part explain this discrepancy in responsiveness to change.
In light of these findings, we recommend that future studies of hypertension should consider using both the MINICHAL instrument and EQ-5D-5L in tandem for the assessment of HRQoL.
Future implications
The availability of an English language, short, validated, disease-specific instrument for the evaluation or HRQoL in hypertension is of value, particularly given the high prevalence of this condition and therefore its wide applicability to patients. Non-adherence to treatment is a crucial element of apparent treatment-resistance in hypertension (42) and therefore the ability to monitor the impact of hypertension and its treatment for patients could help address this important limiter to successful treatment. Furthermore, it is anticipated that the newly-adapted and validated English language instrument will be used in future research practice to ensure that new treatment strategies for hypertension positively impact HRQoL.