Andrew Garratt’s research while affiliated with Norwegian Institute of Public Health and other places

Background: Six to nine per cent of all newborn infants require admission to a neonatal intensive care unit (NICU) due to either illness or prematurity. During their stay, these infants are often subjected to many painful procedures that can cause negative long-term consequences. To reduce the negative effects of pain exposure and ensure optimal and safe pain treatment, accurate assessment of pain is necessary. To achieve this, clinicians are dependent on the use of reliable, objective, and standardised clinical rating scales of pain, henceforth referred to as 'rating scales'. Numerous rating scales have been published; however, discrepancies in validity limit their overall applicability in clinical practice and research. Such limitations may lead to an over- or underestimation of pain, resulting in unnecessary sedation or inadequately treated pain, potentially jeopardising infant safety through treatment side effects, including withdrawal symptoms or prolonged discomfort. To date, the majority of rating scales have been developed to assess procedural pain, whilst fewer scales for prolonged pain are available. Premature infants further complicate matters, as they often have a reduced ability to display robust pain behaviour due to their immaturity. Research has also shown that the use of rating scales in clinical practice is suboptimal, due to both inadequate and infrequent implementation alongside inappropriate choice of scale for the specific pain, population, or setting under evaluation. Despite numerous studies investigating the burden of pain in newborn infants, little work has been done to summarise the current evidence on the appropriateness of rating scales for specific types of pain or infant conditions. This has likely been limited by the subjectivity of pain assessment and further complication of assessing such a non-verbal and immature patient population. The immense burden of neonatal pain worldwide has also led to the development of numerous rating scales in various languages, further hindering evidence summation. Objectives: To systematically review the literature to compile and describe the development, content, and measurement properties of clinical rating scales for the assessment of pain in newborn infants. Search methods: An Information Specialist systematically searched CENTRAL, PubMed, Embase, and CINAHL. The latest update search is current to July 2023. Selection criteria: We included all study designs that involved the development or testing of a rating scale for assessing pain in newborn infants. We included preterm (born before week 37) and term (born at week 37 or beyond) infants undergoing pain assessment for any medical indication. We also included studies that included healthcare professionals. Data collection and analysis: We evaluated clinical rating scales assessing pain in newborn infants using the Consensus-based Standards for the selection of health Measurement Instruments (COSMIN) methodology evaluating content validity, structural validity, internal consistency, reliability, measurement error, hypothesis testing, and cross-cultural validation. We used a modified GRADE approach to assess risk of bias, inconsistency, imprecision, and indirectness. Main results: We included 79 studies involving a total of 7197 infants, 326 nurses, and 12 physicians. Twenty-seven clinical rating scales were used in 26 countries, with 14 studies evaluating preterm infants, 11 on term infants, 46 on both preterm and term infants, four solely on medical staff, and four on preterm and/or term infants plus medical staff. Following the COSMIN checklist, we found all rating scales to be of very low-certainty evidence, raising concerns regarding their validity, reliability, and applicability in this vulnerable population across diverse clinical settings. Authors' conclusions: Clinical staff should be vigilant when applying the currently available neonatal rating scales. Further development of rating scale content and testing for structural validity are necessary and should be prioritised. Together, they determine the content and structure of rating scales, underpin further testing, including reliability, and their prioritisation will make the greatest contribution to the evidence base for rating scales to assess neonatal pain. Collaborative efforts between clinicians and methodology experts will prevent methodological pitfalls and contribute to improving the validity and reliability of pain-rating scales in neonatology.

Background Trauma is a major cause of long-term morbidity and disability. Survivors of trauma may experience negative physical, socioeconomic and quality of life related efects years after injury. Early identifcation of vulnerable patients can help improve tailored care to enhance health and quality of life. PROMs information can be used to identify such patients and monitor their changes in health over time. The PROMIS-29 is a modern short-form health profle that assesses mental, physical and social health across several domains. This study aims to assess feasibility and response rate (RR) of PROMIS-29 as an electronic PROM in trauma patients. Methods Retrospective study including patients>18 years admitted by a multidisciplinary trauma team at St. Olavs University Hospital from 01.09.21 to 31.10.22. Consenting patients were to complete PROMIS-29 at baseline, after 14 days and after 3, 6 and 12 months post-injury. Baseline was completed at the hospital or by smartphone following discharge. Follow-up was completed electronically. Results 196 patients were included. The mean age was 51 years and 34% were women. All patients completed the baseline measurement. At follow-up, the RR was 59% at 14 days, 49% at 3 months, 47% at 6 months, and 41% at 12 months. Conclusion/Discussion The largest reduction occurred between baseline and 14 days post-injury, when patients completed PROMIS-29 electronically. PROMs information is necessary to inform patient care but the downward trend of RRs as seen here and in cohort and registry-based studies, suggests suboptimal methods of data collection. Reductions in RR may stem from patients having problems with the electronic solution, or them being less motivated to respond over time when their reporting does not afect their care plan. Therefore, a follow-up study is planned which will include trauma patients completing PROMIS-29, alongside qualitative interviews. Based on the responses to PROMIS-29, tailored follow-up by a rehabilitation specialist will be ofered.

Purpose To develop the Norwegian value set for the EQ-5D-5L based on interviews with a representative sample of the Norwegian adult population. Methods Random and quota sampling were used to recruit the sample of adults (age> 18 years) representative of the Norwegian general population. Data collection followed EQ-VT 2.1 undertaken before and after the COVID-19 pandemic from November 2019 to December 2022, using PC-assisted and video conferencing interviews, respectively. Each respondent valued 10 health states using composite time trade-off (cTTO) and 7 health states using a discrete choice experiment (DCE). Different statistical models were assessed for logical consistency and predictive accuracy using cTTO and DCE data alone or in combination as hybrid models. Results Of the 1,321 respondents, 1,237 met inclusion criteria. All statistical models demonstrated logical consistency. The weighted hybrid model combining both cTTOand DCE data was preferred and had the highest predictive accuracy. Predicted values ranged from -0.453 to 1, and the dimension of anxiety/depression was the most highly valued by respondents, followed by pain/discomfort, self-care, mobility, and usual activities. These findings are not dissimilar to those for most Western European countries, and regression coefficients are closest to those for other Scandinavian countries. Conclusion This study provides the Norwegian value set for the EQ-5D-5L based on health state values obtained from members of the adult general population in Norway. This is an important contribution to economic evaluation and the broader application ofthe EQ-5D-5L in Norway including clinical and health services research, and quality measurement.

The COVID-19 pandemic had a global impact on daily lives, and this study aimed to assess the effects on broader aspects of health in the general population of Norway. This population-based cohort study assessed changes in health of the Norwegian general population from 3 months before to 9 months during the COVID-19 pandemic. Sampling was based on the results of Norwegian surveys designed for collecting general population norms for health measurement instruments. In December 2019, 12,790 randomly selected adults aged ≥18 years received a postal questionnaire. The 3,200 respondents received a similar follow-up postal questionnaire including the EQ-5D-5L, PROMIS-29 instruments, and questions about respondents having or having had COVID-19. Score changes were compared to estimates for the minimal important change (MIC) and age-related change. Association of instrument change scores with baseline characteristics, health problems, and having had COVID-19 was determined using multivariable linear regression. Of 3101 respondents with unchanged addresses, 2423 (78.1%) responded to the second survey. For all respondents, EQ VAS and PROMIS-29 scores for 6 of 8 domains were slightly poorer (p<0.01) than before COVID-19, and the mean change was below the MIC. In multivariable analyses, the greatest number of poorer outcomes were associated with being female, 18–29 years, or ≥80 years of age (p<0.01); > MIC for ≥ 80 years of age and EQ-5D index, PROMIS-29 physical function and social participation. Respondents who had COVID-19 had poorer outcomes for PROMIS-29 social participation (> MIC). Those reporting COVID-19 in their partner/family and not themselves, had poorer outcomes for PROMIS-29 anxiety and social participation. About 9 months into the COVID-19 pandemic, EQ-5D-5L and PROMIS-29 domain scores showed slightly poorer health in the Norwegian adult general population compared to 1 year earlier in the same respondents. The overall changes were less than expected for age-related change. Relatively poor outcomes defined as important, included general health and social participation for the elderly, and the latter for those having had COVID-19. In conclusion, this study found no evidence for a decline in important aspects of adult general population health in Norway that might be attributed to the pandemic at approximately 9 months.

Background Patient-reported outcome measures (PROMs) are essential for assessing the health of patients with chronic venous disease (CVD). Therefore, we aimed to translate the Aberdeen Varicose Vein Questionnaire into Thai language (AVVQ-Thai) and evaluate its reliability and validity. Minimal clinically important differences (MCID) of the AVVQ-Thai also be estimated. Methods International standards for PROM translation were followed including the forward-backwards translation of the AVVQ. Patients with Clinical-Etiology-Anatomy-Pathophysiology (CEAP) C2-C6 with truncal reflux were prospectively included. Venous interventions were used to treat reflux and varicosities. Patients’ characteristics, venous clinical severity scores (VCSS), EuroQol EQ-5D, and AVVQ-Thai were collected pre- and one-month post-intervention. AVVQ-Thai was also collected one to two weeks after the initial visit by reply-paid postal questionnaire. Results The study included 119 patients (30% C2, 29% C3, 28% C4, 11% C5, and 2% C6). The AVVQ-Thai had good internal consistency with Cronbach’s alpha of 0.783 and moderate reliability with the intraclass correlation coefficient of 0.67 (95%CI: 0.50, 0.79). The AVVQ-Thai was significantly correlated with VCSS and was able to discriminate patients with different levels of health problems as assessed by EQ-5D at both pre-and post-intervention, demonstrating good construct and discriminative validity. The median AVVQ scores improved significantly after intervention from 15.4 (IQR 8.3, 24.2) to 4.2 (IQR 1.3, 8.4) in C2-C3, and 18.9 (IQR 14.1, 25.5) to 7.3 (IQR 4.6, 16.3) in C4-C6. The MCID of the AVVQ was 6.21 on the 0-100 scale, which equates to the level of difference necessary to be clinically meaningful. Conclusions AVVQ-Thai has satisfactory evidence for internal consistency, reliability, validity, and responsiveness to change and is recommended for application in Thailand.

Background Frailty is strongly correlated with mortality in intensive care unit patients, yet routine screening among intensive care patients is rarely performed. The aim of this study is to assess frailty and health‐related quality of life (HRQoL) in patients before intensive care admission and to compare this with outcomes after 3 and 12‐months. The Clinical Frailty Scale and EQ‐5D‐5L will be used to assess frailty and HRQoL, respectively. Methods This is an ongoing, prospective observational study including patients from five Norwegian ICU's. Inclusion criteria are patients aged ≥65 years requiring invasive mechanical ventilation for ≥24 h. The Clinical Frailty Scale and EQ‐5D‐5L are administered at baseline (before critical illness) and at 3‐ and 12‐months post‐inclusion. Additional data collected includes patient characteristics, ICU treatment details, illness severity and mortality. The EQ‐5D‐5L will be compared to Norwegian population norms and assessed for measurement properties. Results Inclusion started July 2022 and will be stopped at 350 patients. The study will be completed in 2025. Conclusion The study will assess the feasibility and measurement properties of the Clinical Frailty Scale and EQ‐5D‐5L in ICU survivors by telephone at long‐term follow‐up study and will give additional information on the frailty and HRQoL of intensive care survivors. Clinical trial registration The study is registered in ClinicalTrials.gov NCT06012942. Protocol version 2.7.1, 19.05.2023.

Background Patient-reported outcome measures (PROMs) are essential for assessing the health of patients with chronic venous disease (CVD). Therefore, we aimed to translate the Aberdeen Varicose Vein Questionnaire into Thai language (AVVQ-Thai) and evaluate its reliability and validity. Minimal clinically important differences (MCID) of the AVVQ-Thai also be estimated. Methods International standards for PROM translation were followed including the forward-backwards translation of the AVVQ. Patients with Clinical-Etiology-Anatomy-Pathophysiology (CEAP) C2-C6 with truncal reflux were prospectively included. Venous interventions were used to treat reflux and varicosities. Patients’ characteristics, venous clinical severity scores (VCSS), EuroQol EQ-5D, and AVVQ-Thai were collected pre- and one-month post-intervention. AVVQ-Thai was also collected one to two weeks after the initial visit by reply-paid postal questionnaire. Results The study included 119 patients (30% C2, 29% C3, 28% C4, 11% C5, and 2% C6). The AVVQ-Thai had good internal consistency with Cronbach’s alpha of 0.783 and moderate reliability with the intraclass correlation coefficient of 0.67 (95%CI: 0.50, 0.79). The AVVQ-Thai was significantly correlated with VCSS and was able to discriminate patients with different levels of health problems as assessed by EQ-5D at both pre-and post-intervention, demonstrating good construct and discriminative validity. The median AVVQ scores improved significantly after intervention from 15.4 (IQR 8.3, 24.2) to 4.2 (IQR 1.3, 8.4) in C2-C3, and 18.9 (IQR 14.1, 25.5) to 7.3 (IQR 4.6, 16.3) in C4-C6. The MCID of the AVVQ was 6.21 on the 0-100 scale, which equates to the level of difference necessary to be clinically meaningful. Conclusions AVVQ-Thai has satisfactory evidence for internal consistency, reliability, validity, and responsiveness to change and is recommended for application in Thailand.

Introduction: Self-perceived health status data is usually collected using patient-reported outcome measures. Information from the patients' perspective is one of the important components in planning person-centred care. The study aimed to compare EQ-5D-5L in survivors after out-of-hospital cardiac arrest (OHCA) with data for Norwegian population controls. Secondary aim included comparing characteristics of respondents and non-respondents from the OHCA population. Methods: In this cross-sectional survey, 714 OHCA survivors received an electronic EQ-5D-5L questionnaire 3-6 months following OHCA. EQ-5D-5L assesses for five dimensions of health (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) with five-point descriptive scales and overall health on a visual analogue scale from 0 (worst) to 100 (best) (EQ VAS). Results are used to calculate the EQ index ranging from À0.59 (worst) to 1 (best). Patient responses were matched for age and sex with existing data from controls, collected through a postal survey (response rate 26%), and compared with Chi-square tests or t-tests as appropriate. Results: Of 784 OHCA survivors, 714 received the EQ-5D-5L, and 445 (62%) responded. Respondents had higher rates of shockable first rhythm and better cerebral performance category scores than the non-respondents. OHCA survivors reported poorer health compared to controls as assessed by EQ-5D-5L dimensions, the EQ index (0.76 ± 0.24 vs 0.82 ± 0.18), and EQ VAS (69 ± 21 vs 79 ± 17), except for the pain/discomfort dimension. Conclusions: Norwegian OHCA survivors reported poorer health than the general population as assessed by the EQ-5D-5L. PROMs use in this population can be used to inform follow-up and health care delivery.

We invited individuals aged above 16 years with a congenital transverse reduction deficiency at and above the wrist born in Norway between 1970 and 2006 to complete the short version of the Disabilities of the Arm, Shoulder and Hand Outcome Measure, the 5-Level EuroQoL-5-Dimension instrument, the RAND 36-Item Short Form Health Survey and a single-item questionnaire on arm function, appearance, pain and prosthesis wear. Of 154 eligible participants, 58 (38%) responded. Their scores were not different from the general population. All had been offered prostheses, and 56 (97%) had been fitted at a median age of 1 year (interquartile range 0–2.8). Of the participants, 37 (64%) were still prosthesis wearers, while 21 (36%) were non-wearers or using gripping devices only. Prosthesis wearers had higher levels of ‘vitality’ as assessed by the RAND-36 and rated their arm appearance higher, but there were no other score differences, indicating that prosthesis rejection is not associated with worse functional outcomes. Level of evidence: III