Article Text
PDF
Abstract
Objective To assess if counselling by an occupational physician (OP) addressing experienced workplace barriers and physical activity integrated as a part of low-back pain (LBP) outpatient treatment influences pain, function and sick leave.
Methods Randomised controlled trial in the secondary healthcare sector with 3 months' follow-up. The participants were LBP patients who, independently of sick-leave status, expressed concerns about the ability to maintain their current job. Patients referred for surgery were excluded. The intervention consisted of two counselling sessions conducted by an OP addressing both workplace barriers and leisure-time physical activity. A workplace visit was performed if required. Pain, function and duration of sick leave due to LBP were primary outcomes.
Results A reduction in bodily pain and improvement in physical function both measured by the 36-item short-form health survey questionnaire in favour of the intervention group was found. The change in pain score was found to be clinically relevant. The risk of sick leave for at least 8 weeks due to LBP was significantly reduced in the intervention group. Two secondary outcomes, Fear Avoidance Beliefs about physical activity and maximum oxygen uptake, supported compliance and adherence to the part of the intervention focusing on enhanced physical activity.
Conclusion Two short counselling sessions by an OP combining advice on meeting workplace barriers and enhancing physical activity had a substantial effect on important prognostic factors for LBP patients with moderate to severe symptoms diagnosed in outpatient rheumatological clinics.
Trial registration Current Controlled Trials ISRCTN13071157
- Randomised controlled intervention trial
- workplace factors
- leisure time physical activity
- low back pain
- function
- sick leave
- epidemiology
- occupational health practice
- back disorders
- intervention studies
- rehabilitation
Statistics from Altmetric.com
- Randomised controlled intervention trial
- workplace factors
- leisure time physical activity
- low back pain
- function
- sick leave
- epidemiology
- occupational health practice
- back disorders
- intervention studies
- rehabilitation
What this paper adds
Extensive multidisciplinary bio-psychosocial interventions have been effective in reducing sick leave among low-back-pain patients in different settings, mostly occupational health services and workers' compensations boards; in both settings, interventions have been aimed exclusively at patients on sick leave.
Only a few of these interventions report lesser pain or enhanced function.
A simple short-time intervention involving two sessions of counselling by an occupational physician and a single workplace visit in about 25% of the cases affects pain function and sick leave in favour of the intervention group.
The finding of changes in maximum oxygen uptake and Fear Avoidance Beliefs Questionnaire-physical activity in favour of the intervention group supports compliance adherence to the intervention, as the issue of physical activity and fear-avoidance beliefs was an important aspect of the counselling session.
A shift in paradigm towards the use of systematic but individual tailored counselling in secondary healthcare with counselling on a specialist level could be useful.
Introduction
The lack of sufficient knowledge concerning primary causative mechanisms for low-back pain (LBP)1–4 and furthermore the lack of effectiveness of physical or organisational ergonomic interventions2 5 6 emphasise the need for secondary prevention of negative consequences of LBP.
Work loss and disability pension among people with LBP are serious events for the individual and expensive for society. Prior to the present study, we performed a register study in the same hospital setting as the present study. In this study, a negative vocational prognosis was found for about 40% of the population.7
In recent decades, several intervention studies focusing on vocational prognosis among people with LBP have been published.8–13 The concepts of the interventions vary from the very simple9 10 to extensive multidisciplinary bio-psychosocial programmes.8 11 12 14 The settings of most studies have been occupational health services,8 12 primary care10 or workers' compensation boards.9 12 One recent study11 was based on secondary healthcare. Simple interventions9 10 and multidisciplinary programmes11 14 15 have both shown an effect on the duration of sick leave and return to work, while a few investigations also suggest an impact on function11 and pain.10 The results are not, however, easily translated into clinical practice because of the heterogeneity of the interventions with regard to setting, populations, content, intensity and the broad ranges of initiatives in the programmes.16 17 Some studies support the claim that physical activity as a part of LBP treatment has a positive effect on pain18 19 and on prevention of future relapses,5 13 but the long-term effect and scope of the effect are discussed.20 As far as we are concerned, general moderate-intensity physical activity has not been a part of the multidisciplinary bio-psychosocial or workplace-orientated programmes.
The overall aim of the present study was to investigate the effectiveness of counselling by an occupational physician (OP), aiming at removing experienced workplace barriers as well as at enhancing physical activity of moderate intensity, on pain, function and sick leave after 3 months.
Methods
Participants and setting
The source population in this randomised controlled trial was patients with LBP referred for specialised evaluation at one of two rheumatological outpatient clinics. Patients were referred from general practitioners or other hospital wards. The inclusion criteria were: age 17–63, paid work employment, willingness to accept a workplace visit if needed, concerns about ability to maintain current job independently of sick leave status and Danish-speaking. Exclusion criteria were: referral for low-back surgery, pregnancy and serious comorbidity. A flow chart based on an assessment of eligibility to the final study population is shown in figure 1.
Flow chart based on an assessment of eligibility for the final study population.*Do not want to inform employer about health problems; in the process of changing job; other illnesses; pregnancy; inhabitant from other region/country; workplace visit not possible (offshore installation).
Interventions
Patients were allocated to either usual care or intervention. Usual care would typically consist of a brief instruction in exercises, or readmission to a general practitioner for further contact with a physiotherapist or chiropractic treatment. Both intervention and control patients were tested by one of three experienced physiotherapists as directed by one of the authors (TM). Balance, trunk-muscle endurance and maximum oxygen uptake (ml O2/min/kg)21 were estimated.22 The baseline test lasted 45 min.
The intervention consisted of (1) an initial counselling session by an OP, (2) a workplace visit if required, (3) a 6-week status interview with focus on compliance and adherence to the plan made together with the OP and (4) a 3-month follow-up concluding counselling session with the OP. The two counselling sessions were carried out by a single specialised OP with broad experience within clinical and workplace-orientated occupational medicine including evaluating patients from all sorts of occupations (LDJ). Prior to the intervention, a pilot study including 50 low-back patients was performed (LDJ, BSC, TM) testing logistic and intervention manual. The two rheumatologists involved (BSC and FHM), two OPs (LDJ and PFr) and the managing physiotherapist (TM) attended a course on dialogue focusing on the patients' perspectives and experienced health concerns.
Initial counselling session with the OP
Results from the baseline test and conclusions from the rheumatological examinations from the day of inclusion were available to the OP.
The aims of the initial counselling session were:
–to establish knowledge about the patient's work situation and leisure-time activity;
–to remove barriers for maintaining the current job and obtaining 45×3 min of moderate-intensity physical activity a week; and
–to establish an individual tailored plan based on the patient's daily working and private life.
The interview and counselling session comprised the following:
–Examination of the patient's usual work situation, including work function, work tasks and general work demand. Further, the patient was asked to describe tasks, postures or situations or situations, if any, that they found strenuous in relation to the actual LBP problem.
–Examination of the actual level of physical activity, qualitatively assessed by asking about activities during a typical week. Emphasis was placed on making the concept of physical activity of moderate-intensity aerobic activity understandable.23
If low-back concerns were present in relation to actual working tasks or physical activity during leisure, such concerns were addressed by having the patient simulate work task or physical activity—for example, a short run on the premises, as part of the counselling session, or by making references to the physical test just accomplished. If the patient experienced social problems, such as worker's compensation or job insecurity, they were given advice on how to manage these issues.
Based on the level of LBP problems, workplace barriers experienced and physical activity, one to three goals for the following 3 months were stated in cooperation with the patient. The goals concerning the work-site factors could include a scheme for a gradual return to the patient's usual work, temporarily avoiding tasks that are experienced to be strenuous for the lower back, or possibilities for more variation during the working day.
The goal for physical activities was generally 3×45 min a week of moderate intensity with a detailed tailored plan for the following 3 months. The plan could be bicycling to work once a week, running one afternoon after work, swimming or playing football with the kids in the weekend. The patients were encouraged to proactively use the follow-up period to experiment with their work method, tasks and physical activity. The patients were informed that relapses were to be expected.
At the first counselling session, a pamphlet describing the project rationale and steps in the investigation was given to the patient, with a copy to be delivered to their supervisor.
The initial OP counselling lasted between 45 min and 1 h, and a workplace visit lasted 1–1½ h.
Workplace visit
A workplace visit was arranged if it was not possible for the OP to gain sufficient comprehension of the work situation from the initial counselling session. The workplace visit was arranged in cooperation with the patient. At the workplace visit, the OP, the supervisor in charge of the patients working conditions and the patient would be present. The usual tasks were overseen, and solutions for any barriers were discussed. In the cases where a workplace visit was performed, the one to three goals for the following 3 months were determined together with the patient at the day of the workplace visit.
Six-week status interview
Six weeks after the initial counselling session with the OP, a midway interview with the patient was performed by an independent research associate (BG). The aim of this semistructured interview was to monitor compliance and adherence to the goals made at the first counselling. The midway interview lasted about 45 min.
Follow-up counselling with the OP
At the 3-month counselling session, the experiences in the follow-up period were evaluated, and a final conclusion concerning vocational future and level of physical activity was made in cooperation with the patient. The follow-up counselling session with the OP lasted 45–60 min.
Outcome measures
Primary outcomes
Primary outcomes were changes in back-specific and generic measures of pain and function, the proportion of patients accumulating 8 weeks of sick leave and the duration of sick leave among those having sick leave. Pain and function were assessed by questionnaires administered by the project nurse at baseline and mailed after 3 months. Low-back-specific pain was assessed using an 11-point numerical rating scale (NRS) scoring of mean pain during the last 3 months. Back-specific function was assessed using the Roland Morris Disability Questionnaire.24 Generic health status was assessed using the 36-item short-form health-survey questionnaire (SF-36) physical function (PF) and bodily pain (BP).25–27 Sick leave was assessed in two independent ways: (1) self-assessed number of sick leave days due to LBP in the 3-month follow-up period and (2) register-based sick leave due to all causes based on data from the Danish National Register on Public Transfer Payments (Danish acronym DREAM)28 for the same period. The DREAM register includes week-by-week records of any type of public transfer payment at an individual level. Sick-leave periods of less than two consecutive weeks are not reimbursed from public sources and are not recorded in the DREAM register.
Explanatory outcomes
Maximum oxygen uptake (ml O2/min/kg)21 and fear avoidance assessed using the Fear Avoidance Beliefs Questionnaire (FABQ) were both measured at baseline and follow-up.29 The FABQ consist of two parts: FABQ-work (W), in this study five items, and FABQ-physical activity (PA), four items. At the status interview after 6 weeks performed in the intervention group, the level of adherence to the intervention goals was evaluated on a 0–10-point NRS anchored with zero=‘the goal not reached at all’ and 10=‘the goal reached to a high degree.’ The interviewer repeated the wording of each of the goals from the initial counselling session with the OP and asked: ‘On a scale from 0 to 10 to what extent are the goals met?’ This rating was the basis for subsequent elaborating questions with focus on barriers and supporting factors concerning compliance and adherence to the goals.
Sample size
We expected a mean reduction in pain score of 2.5 on an NRS as a natural reduction in pain over 3 months in both the intervention and control groups. A between-group difference from baseline to follow-up of 1.0 (SD 2.4) was anticipated as a relevant difference between the groups. To detect this difference with power of 0.90 at the 0.05 level of significance, 122 subjects were needed in each group. As we expected a follow-up rate of around 80%, 150 subjects were included in each group.
Randomisation
The source population included all LBP patients from the two outpatient clinics during the study period. The patients came from a large admission area, so the likelihood of seeing two or more patients from the same working place was small. The RANNOR computer algorithm was used for the single randomisation. A person with no association with the allocation sites prepared sealed envelopes. A project nurse, who was not involved in diagnosing the patient, evaluated whether the patient met the inclusion criteria. Oral and written information was given to those who met the inclusion criteria. When written informed consent was obtained, the project nurse would open a sealed envelope for random assignment to either group.
Blinding
The assignment of group was not blinded for the physiotherapist performing the baseline physical tests or to the OP, as the result of the physical test was an integrated part of the counselling by the occupational physician.
Statistical analyses
Descriptive statistics were used to evaluate baseline characteristics in the two groups: differences in age, gender, vocational history and Quebec-task-force-classification.30
Differences among dropouts and participants were evaluated using the Student t test and χ2 test. The means of between-groups differences from baseline to follow-up of pain, function and explanatory outcomes were evaluated using paired analyses. Differences in between-groups means were calculated together with their 95% CI. An improvement of at least 30% in scores from baseline was considered a clinically meaningful improvement for a range of LBP-relevant outcome measures.31–33 The relative numbers of patients obtaining clinically meaningful improvement in pain and function in the intervention group and control group were compared using the χ2 test. Relative risks and 95% CI of accumulating at least 8 weeks of sick leave due to LBP as well as at least 8 weeks of register-based general absence were estimated together with their 95% CI. Per protocol and intention-to-treat analyses were performed for register-based sick leave. Among patients on sick leave, the total duration of self-reported sick leave related to LBP as well as register-based general sick leave was analysed using the Cox Proportional Hazard model including intervention group as the only independent variable yielding HRs together with their 95% CI.34
Results
The number of LBP patients seen in the two outpatient clinics during the study period was 1461. A flow chart addressing eligible patients, randomisation and course of the study is shown in figure 1. Of the 300 patients included, a total of 224 patients, 110 from the intervention group and 114 from the control group, completed the 3-month follow-up analyses and thus comprised the final study population with a follow-up response of 73% respectively 76% for the intervention and control group. Inclusion took place from November 2006 to April 2009. Follow-up was finished in July 2009, and register data including the whole population were available 6 months later. The mean time from inclusion to follow-up was 104 calendar days in the intervention group and 102 in the control group. As shown in table 1, 25.5% of the patients were professionals and/or held university degrees or similar, 53.6% were in teaching, office work or nursing, while 20.9% were blue-collar workers, a distribution similar to the population in the newly published population-based Dutch study.11 The original study plan included a workplace visit as part of the intervention to support the intervention or to determine the working conditions. Workplace visits took place in only 29 (26%) cases, mainly at blue-collar workplaces, as workplace barriers in the more sedentary jobs in most cases could be dealt with as part of the first counselling session—with the OP. Workplace visits were arranged within 2 weeks from the first counselling session and did not extend the duration of the intervention period. The time consumption in the treatment system beside the usual investigation by the rheumatologist was 8 h for patients with a workplace visit and 5 h for patients without a workplace visit.
Baseline characteristics of the study population
Loss to follow-up
There were no significant differences between allocated, baseline and follow-up populations concerning, age, gender and vocational history from 2 years before inclusion, and Quebec-task-force Classification of clinical characteristics (results not shown).
Baseline characteristics
Table 1 presents the baseline characteristics. No differences were observed in the primary outcomes or in the possible prognostic factors between the groups. A slightly higher level of FABQ-W was found in the control group. More than 70% of the study population had radiating pain, and the pain score was about 6 in both the intervention and control group.
Primary outcomes
Both generic measures from SF-36 improved significantly in favour of the intervention group with a mean between-group difference in BP of 6.27 (95% CI 0.70 to11.83) and PF of 6.60 (95% CI 1.39 to 9.81). The mean between-group difference of back-specific pain was 0.66 (95% CI 0.12 to 1.45) and back-specific function 1.0 (CI −0.32 to 2.33); both were insignificant although in the same direction as the generic measures. Both measures of pain were found to be clinically meaningful: a statistically higher proportion of patients in the intervention group lowered their pain with more than 30% compared with the control group.31 The mean between-group differences for the back-specific and generic outcomes are shown in table 2. Table 3 summarises the findings regarding sick leave, including intention-to-treat analyses performed on the register-based sick-leave data. The risk of sick leave for at least 8 weeks due to LBP was reduced in the intervention group, as was the register-based sick leave due to all causes. Among those on sick leave due to LBP, the duration of such sick leave was greater in the control group, yielding an HR of 2.57 (95% CI 1.52 to 4.37). The cumulated number of working days on sick leave because of LBP was 406 in the intervention group versus 1081 in the control group. Among those who had register-based sick leave due to any cause, the effect was less pronounced (HR 1.24 (95% CI 0.77 to 2.00), with 1595 lost working days in the control group versus 1100 in the intervention group. None of the participants changed to disability pension or other types of permanent transfer income during the follow-up period.
Between-group differences in primary and explanatory outcome measures from baseline to follow-up
Risk estimates of self-assessed sick leave because of low-back pain and register-based sick leaves of all causes
Explanatory outcomes
As shown in table 2, both the intervention and control groups increased their maximum oxygen uptake during the follow-up period, but the between-group difference was statistically significantly elevated in favour of the intervention group with 1.6 ml O2/min/kg (95% CI 0.06 to 3.20). FABQ-PA changed significantly with a between-group difference of 3.10 (1.51 to 4.69), while no changes were found in FABQ-W. The mean (SD) of the self-assessed goal accomplishment was 5.8 (3.4) on a 0–10-point NRS.
Only 4% of the patients had goals concerning a substantial change in physical working conditions, 55% had goals concerning a greater variation or pauses during the working day, and 40% had goals concerning temporary changes to their working tasks or working time. Greater physical activity was a goal for 80% of the patients.
Discussion
In this randomised controlled trial, two counselling sessions by an OP addressing experienced workplace barriers, and moderate-intensity physical activity was more effective than usual care in reducing SF-36 BP, enhancing physical function measured by SF-36 PF and maximum oxygen uptake, and reducing the amount of self-reported sick leave. No significant differences were found in the low-back-specific pain measures and Roland Morris Disability Questionnaire. The reduction in BP and pain intensity was found to be clinically meaningful as well.31 We cannot determine if one element of the intervention is more important, as the counselling functioned as a whole. A reason for the adherence to the goals and the significant increase in maximal oxygen uptake could be the emphasis on the patient's perspective, that is the counselling targeting elements the patient themself expressed as barriers and the very specific plans for physical activity.
The finding of significant effect sizes in favour of the intervention group in generic measures, but not in the back-specific measures, could be a result of the substantial focus on any kind of general physical activity both in private life and at the workplace during the OP counselling session.
The intervention group had less than half the number of self-assessed sick leave days compared with the control group and a significantly lower proportion of persons on sick leave for more than 8 weeks during the 3-month follow-up period. Among those who had register-based sick leave, the effect was less pronounced. The difference between self-assessed sick leave and the register-based data could be due to recall bias and the fact that the register-based data included sick leave due to all causes, which would tend to dilute the differences caused by LBP between the intervention and control group.
Sick leave was not an inclusion criterion, as sick leave can be unevenly distributed across occupations35 and so would violate the generalisability to populations with different occupational profiles from the actual investigation. The reasons for this uneven distribution could be different physical demands in different occupations, differences in welfare payments and more restrictive counselling from healthcare professionals to patients in occupations thought to be heavy. Furthermore, it can be difficult to compare sick-leave measures across countries and prevailing economic conditions, as legislation, insurance systems and incentives for work differ. In Denmark, sick leave is restricted to 52 weeks, which can mean that patients who face a transition to a lower transfer income are more motivated to return to work despite persistent symptoms, compared with patients in other countries with longer sick-leave benefit periods.
In contrast to our expectations when planning the study, a workplace visit was required in only about 25% of all cases. The reason for the limited need for a workplace visit could be the demographic of this population-based study, which had only about 20% blue-collar workers, as well as an overall tendency towards less physically demanding working tasks in Denmark and other industrialised countries.
Limitations and strengths of the study
A main limitation was the lack of blinding of both assessor and patient. The test and retest by a physiotherapist were introduced to ensure that control patients did not feel that they missed out on treatment. Assessment bias in maximum oxygen uptake would be minimal, as it is an objective measurement calculated from parameters where judgement from neither the assessor nor the patient is of importance: pulse, workload and body weight. The patient's knowledge of assignment did not affect the completeness of follow-up, but we do not know the degree to which their perception of symptoms has been affected. We do not know if there is an underestimation of the part of the effect related to physical activity, as the control group was also tested, meaning that they, too, could be motivated to enhance their level of physical activity like the intervention group. As the study evaluates a combination of experienced workplace barriers and encouragement of general physical activity, it is not possible to evaluate the two parts separately. The randomisation was successful, and no contamination took place, as all patients came from different workplaces. As regards external validity, our population was strictly defined by setting, clinical and demographic characteristics. The source population was LBP outpatients referred to a rheumatologist either at a university hospital or at a rheumatological clinic with a pain score of about 6, 70% of the patients had radiating symptoms, and about 11% had more than 90 days of sick leave during the year before inclusion. The outpatient clinics were covering the same mixed urban and suburban geographical area. In Denmark, which has a population of about 5 million people, approximately 670 000 persons come into contact with their general practitioner because of LBP each year. About 5% are referred to secondary care at an outpatient clinic.36 Denmark has a public, free healthcare system that minimises socio-economic differences in access to care. Our population includes patients from a variety of social strata, comprising 27% highly educated professionals or managers, 51% public servants such as teachers, office workers and nurses, and 22% blue-collar workers.
As our source population comprised an unselected population, the findings concerning pain and function from this study can be generalised to LBP patients not referred to surgery, having a workplace connection, and sharing the clinical and psychosocial characteristics of the actual study population. Caution must be exercised when generalising results concerning sick leave to countries with different social legislation.
A common difficulty in intervention studies is a lack of knowledge of compliance and degree of adherence to the intervention programme. In this study, the assessments of maximum oxygen uptake and FABQ as explanatory outcomes together with results from a status interview after 6 weeks contribute information on changes in behaviour possibly initiated by the intervention.
Comparisons with other studies
Our finding of a clinically relevant effect on pain contrasts with the majority of multidisciplinary bio-psychosocial interventions.8 11 12 37 38 Two recent studies39 40 stress the importance of pain as a predictor of chronicity, especially a clinically relevant decrease in pain intensity during the first 3 months.40 Only a few intervention studies demonstrate positive findings in function among LBP patients. Our finding of an improvement in SF-36 PF is supported by the finding of an improvement in maximum oxygen uptake as well. A recent study11 investigating the effect of integrated care found differences in functional status at a 12-month, but not 3-month, follow-up. The finding of a reduction in sick leave during the intervention period is in line with both the majority of the extensive bio-psychosocial interventions and the more simple concepts.8 9 11 37 In this intervention study, we focus on disability management at the individual level with a combination of dealing with experienced workplace barriers and encouraging general physical activity in the overall context of lowering concerns of LBP disability in line with the multidisciplinary bio-psychosocial context.14 15 In contrast to the multidisciplinary bio-psychosocial studies, our study is based on just two counselling sessions by an OP and a workplace visit if required, with all advice given by one person, that is, the OP. In these respects, our study is in line with the minimalist concept introduced by Indahl9 and a later Finnish study.10 As in those two studies, we include a physical examination and conclusion by a specialist, information regarding the—usually good—prognosis, and encouragement to remain active. Compared with those studies, we have added:
a physical test at the start of the study in order to have a point of reference for advice concerning general physical activity and in order to address low-back concerns if necessary;
combined counselling including attempts to remove workplace barriers and advice to obtain moderate-intensity physical activity;
a status interview after 6 weeks to evaluate comprehension and adherence to the advice given; and
encouragement to proactive use of the follow-up period to experiment with work and physical activity—experiences were evaluated at the mandatory follow-up interview.
Policy implications
As there is limited knowledge concerning primary causative mechanisms for LBP1–4 and results from physical or organisational ergonomic interventions,2 5 6 secondary prevention is the most realistic option. We propose that the target population for prevention of serious consequences of LBP is in the secondary healthcare sector, as most onset cases in the primary sector have a benign self-limiting cause. The use of a tailored combined counselling with emphasis on experienced work-site barriers and moderate-intensity physical activity for a minimum of 3×45 min a week integrated in secondary healthcare with counselling on specialist level would have several advantages (compared with later counselling when the patient is returned to the primary sector or the social security system). The gain of this approach would be substantial for patients and society alike, saving on sick-leave benefit, further treatment costs and case management in the social security system. From the present study, it is estimated that this counselling would be relevant for about 30% of LBP patients seen for specialised evaluation in a secondary healthcare setting.
Conclusion
In summary, this study supports that occupational counselling based on two counselling sessions by an OP addressing experienced workplace barriers and physical activity integrated in LBP secondary care has an effect on prognostic factors for LBP inpatients seen in outpatient clinics with moderate to severe symptoms. Our findings suggest that there should be a shift in paradigm with more attention on a systematic but individual tailored counselling as part of the concluding treatment for LBP outpatients in the secondary sector. The counselling should include physical test a counselling session addressing experienced workplace barriers and physical activity and a three month follow up. The results concerning changes in maximum oxygen uptake are to be corroborated in studies designed for that purpose.
References
Footnotes
See Editorial, p 1
Funding Danish Research Fund for the Working Environment. Grant number 10-2005-09.
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics approval was provided by the Danish Data Protection Agency, number 2006-41-6190, and notified to the Central Denmark Region Committees on Biomedical Research Ethics.
Provenance and peer review Not commissioned; externally peer reviewed.