Assessment of the performance of the Brazilian Portuguese Nottingham Health Profile in adult growth hormone deficiency and pulmonary hypertension

Introduction

The Nottingham Health Profile (NHP) is a generic measure of perceived distress that has been used widely as an outcome measure in clinical practice and trials. It consists of a profile of outcomes assessing different types of distress: physical mobility, pain, sleep, social isolation, emotional reactions, and energy level¹. Generic measures are considered valuable as they enable comparison of patient-reported outcomes between different populations and in assessments between healthy and unwell populations². The NHP has also been widely used as a comparator measure when validating disease-specific patient-reported outcome measures (PROMs). Newer, disease-specific PROMs have the advantage of asking questions that are more relevant to patients, while omitting questions that are less relevant.

As a result of the lack of responsiveness and age of the generic measures, they are being replaced by disease-specific PROMs. However, in medium and low-income countries, there is a lack of comprehensive information systems, creating methodological obstacles in evaluation studies and limiting the capacity to conduct longitudinal studies³. Healthcare professionals, researchers and pharmaceutical companies are therefore reliant on generic outcome measures to collect information from patients because, theoretically, they can be used with any patient population. However, this means that outcome measures are selected based on availability rather than on quality.

The NHP has been used in two recent Brazilian studies. The measure was administered to patients with adult growth hormone deficiency (GHD) and pulmonary hypertension (PH) as comparator measures in the evaluation of Brazilian adaptations of the Quality of Life Assessment of Growth Hormone Deficiency in Adults (QoL-AGHDA) for GHD and the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) for PH^4,5. The availability of these data sets for secondary analysis provided the opportunity to assess the psychometric performance of the NHP in two different Brazilian populations.

Adult GHD results from decreased growth hormone secretion from the anterior pituitary gland. This is characterized by decreased lean body mass and increased fat mass, hyperlipidemia, cardiac dysfunction, decreased fibrinolysis and premature atherosclerosis, decreased muscle strength and exercise capacity, decreased bone mineral density, and decreased insulin resistance⁶. The most frequent cause of childhood onset is idiopathic and may not necessarily be associated with other pituitary hormone deficiencies. In contrast, adult onset GHD results from hypothalamic-pituitary tumors and/or their treatment⁷. GHD patients present with lower energy levels and more emotional problems than healthy individuals⁸.

Pulmonary hypertension (PH) is a condition that may occur with a variety of disorders and is characterized by an increase in the pulmonary vascular resistance and in the mean pulmonary arterial pressure (mPAP). PH is currently classified into different groups according to similarities in pathophysiology and the presence of associated conditions⁹. Group 1 (pulmonary arterial hypertension (PAH)) and Group 4 (PH due to pulmonary artery obstructions) are the most extensively studied and those for which there are approved drugs and/or procedures. Chronic thromboembolic pulmonary hypertension (CPTEH) is the most prevalent of the latter group. Patients with PH present with non-specific symptoms such as shortness of breath, progressive exertional dyspnea, chest pain, fatigue or syncope that progress over time, leading to right ventricular dysfunction and death¹⁰. PH presents a significant impact on patients’ social and emotional well-being and on their daily life in general, with restrictions in the ability to perform everyday tasks¹¹.

The aim of this study was to investigate the psychometric properties of the NHP in adult GHD and PH populations. The purpose of the new analyses was to see how valuable the NHP could be in assessing outcomes in diseases where there is no effective disease-specific measure available.

Methods

Results

Demographic information for the GHD and PH samples is presented in Table 1. The GHD study population comprised primarily of patients with adult onset GHD. The PH sample consisted of patients who had been diagnosed as Group 1 PAH (n = 62) and Group 4 CTEPH (n = 40). Although the disease groups were not matched, the groups were similar in age. Similarly, both samples consisted of more females than males. Most of the respondents in both samples were married or living as married. However, considerably more GHD patients were employed and fewer were retired, compared to PH patients.

Descriptive statistics for the NHP sections are shown in Table 2. The distribution of scores for each section suggest that the patient groups experienced similar levels of distress. GHD patients reported marginally worse scores on the pain, emotional reactions and social isolation sections, compared to PH patients. However, PH patients scored higher on the physical mobility section. Major ceiling effects were observed for the energy section in both samples. Substantial floor effects were observed for all NHP sections with the exceptions of emotional reactions in the GHD sample and physical mobility in the PH sample. Raw NHP data for each participant are available as Underlying data²².

Cronbach’s alpha coefficients for the NHP sections are presented in Table 3. Coefficients above the minimum acceptable level of 0.7 were obtained for all sections in both samples. Spearman’s rank correlation coefficients between scores obtained on Time 1 and Time 2 for the NHP sections are shown in Table 4. All correlations were significant at the p<.01 level.

Disease duration and perceived general health are presented in Table 5. A Mann-Whitney U-test revealed that the PH sample had significantly worse ratings of general health than the GHD sample (U = 5070.5, p < .05).

Table 6a and Table 6b show NHP section scores grouped by perceived general health in the GHD and PH samples, respectively. For both samples, most of the NHP sections were able to distinguish between groups of respondents that differed according to their perceived general health. For these NHP sections, individuals who considered their general health to be ‘Excellent / Good’ had significantly better scores on the NHP sections than those who rated their health less favourably. The exceptions were the sleep section in the GHD sample and the social isolation section in the PH sample.

Table 7 shows NHP section scores by age group (above versus below median age) and gender in the adult GHD and PH sample combined. A Mann Whitney U-test found no significant difference in scores between older and younger patients on any of the NHP sections. Significant differences in scores were found between females and males. Females reported significantly higher scores than males on all sections except energy level and social isolation.

Figure 1 shows the Spearman’s rank correlations between scores on the NHP sections and disease-specific measures of QoL in patients with GHD (QoL-AGHDA) and PH (CAMPHOR QoL scale). Total scores on the QoL-AGHDA correlated most highly with the emotional reactions and energy level sections. The CAMPHOR QoL scale correlated relatively highly with the emotional reactions and social isolation sections but was also influenced by problems with physical mobility. A weak association was observed between total scores on the QoL-AGHDA and the sleep section of the NHP. Scores on the pain section of the NHP did not correlate highly with QoL scores on either the QoL-AGHDA or the CAMPHOR.

Figure 1. Spearman’s rank correlation coefficients between NHP section scores and total scores on the QoL-AGHDA and CAMPHOR QoL scale.

Note: All correlations were significant at p<.01.

Discussion

The use of patient-reported outcomes to evaluate different interventions is becoming commonplace in clinical studies and trials. For many years, generic outcome measures such as the SF-36 and NHP were used for this purpose. However, these generic measures have been shown to lack sensitivity to change over time. In addition, all generic measures lack the responsiveness to be fully effective in clinical trials because their items are intended to be suitable for all possible illnesses. This means that they miss important aspects of the disease and include irrelevant items^23–25. Consequently, there has been a move towards the use of disease-specific outcome measures which are able to ask more relevant questions and to measure outcome more accurately. Many researchers also continue to use the generic outcome measures together with disease-specific measures. While disease-specific measures are becoming more widely available, they have yet to be widely adapted for use in medium- and low-income countries. At the same time, clinical trials are being more frequently undertaken in such countries. Such studies require patient-reported outcome measures.

The current study looked at the performance of the generic NHP in two patient groups in Brazil, adult GHD and PH. NHP data were available from these patients as the measure had been used to help validate new disease-specific outcome measures. Analyses were undertaken to see whether the NHP would be a useful outcome measure for trials in other diseases.

The two populations had comparable ages. However, a greater proportion of PH patients were retired. This difference could be explained by the PH sample experiencing poorer general health, compared to the GHD sample. NHP section scores in both groups were strongly related to perceived general health.

Scores on the NHP sections were not influenced by age but women scored higher (had worse health status) than men. These findings are consistent with previous research reporting gender differences in QoL impairments for both patient populations studied^26,27.

Overall, the psychometric properties of the NHP sections were good for both disease populations. Estimates of internal consistency suggest that the items in the six sections of the NHP are sufficiently inter-related to form scales. Test-retest reliability (reproducibility) was better than that achieved by other generic measures though slightly lower than ideal. While only pain and emotional reactions demonstrated test-retest reliability coefficients of 0.85 and above in the GHD sample, all NHP sections in both patient groups were above the usually quoted acceptable value of 0.7^28,29.

A large proportion of GHD and PH patients obtained the lowest possible score on most of the NHP sections. This indicates that the measure is not well targeted to these samples, which could be explained by the relatively mild perceived general health of the GHD and PH patients. Scores on the NHP confirmed the presence of common problems experienced by patients. Both samples reported high scores on energy level, the PH sample scored high on physical functioning and the GHD sample on emotional reactions. The correlations between scores obtained on the disease-specific measures of QoL and the NHP sections suggests that pain and sleep did not seem to be important predictors of QoL in either of the samples.

A limitation of the current study is the lack of clinical information about the patient groups. It must be noted that the data collected for disease duration in the adult GHD sample should be interpreted with caution. The time between initial onset of GHD and diagnosis is likely to affect disease severity. In addition, the lack of inclusion of patients with GHD of childhood onset or GH-treated patients is a weakness, although relatively few GHD patients in Brazil are currently prescribed replacement growth hormone. It would have been valuable to collect WHO functional class information for the PH patients to explore whether the NHP could detect differences in health status relating to objective disease severity.

The use of the NHP in adult GHD and PH populations in Brazil is not recommended as there are high-quality disease-specific measures available for each disease. However, where no disease-specific measures are available, the NHP can provide good descriptive information of the impact of disease on different patient populations.

Data availability