Completeness and Mismatch of Patient-Important Outcomes After Trauma : Annals of Surgery Open

Secondary Logo

Journal Logo

Original Study

Completeness and Mismatch of Patient-Important Outcomes After Trauma

James, Arthur MD, MSc*,†,‡; Ravaud, Philippe MD, PhD*,‡; Riveros, Carolina MSc*; Raux, Mathieu MD, PhD§; Tran, Viet-Thi MD, PhD*,‡

Author Information
Annals of Surgery Open: December 2022 - Volume 3 - Issue 4 - p e211
doi: 10.1097/AS9.0000000000000211

Abstract

INTRODUCTION

An outcome is an event or measure in study participants that is used to assess the effectiveness and safety of an intervention or to measure the health of a population.1 Research stakeholders have been encouraged to build evidence based on patient-important outcomes (PIO).2 Despite a slow increase in the proportion of randomized controlled trials (RCTs) measuring at least 1 PIO (from 18% of the RCTs in diabetes in 2007 to 24% of RCTs including critically ill patients in 20173–5), most studies still do not collect outcomes that matter to patients.

Furthermore, measuring a single PIO may insufficiently capture the effect of an intervention or to measure the health status of a population in most conditions. For example, injuries, which represent about 255 million disability-adjusted life years (DALYs) lost every year,6 and among them, severe trauma, until recently have been almost exclusively evaluated with short-term outcomes such as in-hospital mortality.7,8 Yet, with now 80% to 90% of severe trauma victims surviving in France9–11 and over the world,12–15 severe trauma has become a long-term condition involving multiple and complex consequences on patients’ lives. A recent qualitative study assessing patients’ perspective on important domains that should be assessed after trauma found 97 important issues ranging from body dysfunction (eg, pain, joint stiffness, or neurological disabilities) and psychological suffering, to broader life consequences such as the burden of treatment (eg, having to attend regular medical tests and consultations years after the trauma).16

To transform the assessment of severe trauma and include the collection of long-term patient outcomes that matter to patients, researchers need to decide the most important outcomes to be collected in trials.17 One response is the use of Core Outcome Sets (COS), which consist of a short list of clinical outcomes deemed most important to patients, care providers and researchers within a specific clinical area.18 COS allow, within a consensually and a priori defined framework, to systematically orient research stakeholders towards a set of standardized important outcomes. Yet, with rare exceptions, COS uptake in trials is limited.19 Furthermore, among existing COS, less than 5% involved patients in their development despite evidence that failing to include the views of end users might overlooked important outcomes and result in research waste.20–23

We would like to argue that the current metrics used to measure the collection of PIO in clinical research might overlook several issues. First, many outcomes not included in a COS might still be important to patients, and simply assessing COS uptake does not provide any information about their collection. Second, binary metrics (ie, assessing least 1 PIO or not; or measuring all outcomes from the COS or not) overestimate or underestimate the completeness of the collection of patient-important outcomes in a study or in a field (ie, a study measuring 4 out of 5 outcomes included in the COS is not considered as using the COS). Finally, even though studies in a field might measure several PIO, some domains might be systematically overlooked, resulting in a mismatch between patients’ perspectives and outcomes measured in clinical research.

Taking the example of severe trauma, this study aimed to identify all outcomes used to measure health at least 6 months after severe trauma in order to provide a comprehensive view of (1) the completeness of the collection of patient-important outcomes in study manuscripts and directly from patients; and (2) the mismatch between outcomes measured and patient-important issues after trauma.

METHODS

In this study, we first performed a systematic review of all outcomes used in clinical studies evaluating long-term outcomes after trauma. Second, we compared the measurement tools used in these outcomes with the results of a previously published qualitative study, exploring patients’ perspective on important issues after trauma. This allowed us to assess (1) the completeness of the collection of patient-important outcomes at study and field level; and (2) the mismatch between outcomes measured and the issues elicited by patients.16

Definitions

Outcomes

Outcomes are events or measures in study participants that is used to assess the effectiveness and safety of an intervention or to measure the health of a population. According to Zarin et al,24 the complete and precise description of an outcome includes (1) its domain (eg, anxiety), its specific measurement (eg, Beck Anxiety inventory), its specific metric (e.g, end value), its method of aggregation (eg, the proportion of patients with an improvement >50%) and its timeframe. This description is similar to the one recommended in the SPIRIT guidelines.25

Domains

Outcome domains are meant to encompass the broad purpose of each outcome. In our previous example, the outcome domain was “Anxiety.”

Measurement Tools

Measurement tools are the specific instruments used to collect outcomes. They can be simple questions, scales or clinical examination elements, biological parameters, etc. In our previous example, the measurement tool was “Beck Anxiety inventory.”

Step 1: Systematic Review of Outcomes Used in Clinical Studies Evaluating Long-term Outcomes After Trauma

Eligibility Criteria

We identified studies involving adult patients (≥18 years) that collected at least 1 outcome after at least 180 days (ie, 6 months) after a severe trauma. Severe trauma was defined as injuries that involved (1) at least 2 body area of the Abbreviated Injury Score (AIS) system,26,27 or (2) an Injury Severity Score (ISS) ≥15,27 or (3) brain injuries qualified as moderate to severe based on structural imaging, length of loss of consciousness, post trauma amnesia, Glasgow Coma Score, and the head Abbreviated Injury Score (AIS).28 We made the methodological choice to include TBI because many severe trauma also have TBI and many TBI patients also have severe injuries, with an overlap of these 22 populations. We excluded studies that focused only on burn patients. We included clinical trials (controlled or not; and randomized or not) and observational studies (cohort, register, cross-sectional, longitudinal). Systematic and nonsystematic reviews, diagnostic studies, editorials, abstracts, posters congress, case reports, and articles for which full texts were unavailable were not included. We included only reports published in English, French, German, or Spanish.

Information Sources and Search Strategy

We searched for eligible studies in Medline (via PubMed) and EMBASE from January 1, 2014, to September 30, 2019. We also searched Clinicaltrials.gov for ongoing or unpublished studies using the same date range. The electronic search used both controlled vocabulary and a wide range of free-text terms (Supplemental Material 1, see https://links.lww.com/AOSO/A174).

Study Selection

After removing duplicates, 2 authors (AJ, CR) independently examined the title and abstract of each reference to identify relevant studies. As many studies did not provide a clear definition of severe trauma and used broad terms such as “multiple trauma,” “severe injuries,” or “polytrauma”; 2 authors (AJ, VTT) decided whether they should be included or not based on the reported baseline ISS, AIS, and proportion of severe trauma brain injuries. Disagreements during study selection were discussed by the authors (AJ, CR, VTT) to reach consensus and reasons for exclusion were listed. We used the Rayyan systematic review software to organize the screening process (http://rayyan.qcri.org).

Data Extraction

A data extraction form was designed and pilot tested independently by 2 reviewers (AJ, CR) with 10 studies. One reviewer (AJ) then extracted the data from all studies. Data extraction covered (1) general characteristics of the study (e.g, year of publication, funding source, geographical area, number of centers involved, number of participants included and study design); (2) the characteristics of patients involved for published studies (eg, age, sex ratio, ISS); or the inclusion criteria planned for entries in Clinicaltrials.gov; and (3) outcome details.

For each study, we extracted data on all outcomes reported. We defined as primary outcomes those reported as such by the authors or, if not specified, the first outcome reported in the study. All other outcomes were considered as secondary. For each outcome, we extracted its domain, specific measurement, metric, method of aggregation and time frame, following the framework from Zarin et al.24 For time frame, as outcomes were often measured repeatedly, we extracted the latest assessment time. We classified the outcome using the 12-category taxonomy from Clarke and Williamson29 (eg, Pain, Function, Mental health or Health-Related Quality of Life [HRQoL], Supplemental Material 2, see https://links.lww.com/AOSO/A174). We identified the source of outcome data as clinician-reported, patient-reported, or administrative data.30

Step 2: Comparison Between the Measurement Tools Identified During the Systematic Review and Patients’ Perspective About Important Domains After Trauma

Identification of Domains That Matter to Trauma Survivors

Domains that matter to severe trauma survivors were identified using a previously published qualitative study.16 In this study, 340 severe trauma survivors were included, randomly sampled from a national register covering of all severe trauma patients cared for in level-1 trauma centers of the Paris region. During semistructured telephone interviews, patients reported 97 issues that really matter in their daily life after trauma (Supplemental Material 3, see https://links.lww.com/AOSO/A174). In this previous study, participants’ median age was 41 years [27–54], 266 (78%) were male and 231 (68%) had no comorbidities before the accident. The median ISS was 17 [11–22] with 20% (n = 68) with ISS ≥25. Most participants (n = 223, 66%) required emergency surgery within the first 24 hours of care. At hospital arrival, 10% (n = 34) of participants had hemorrhagic shock and 30% (n = 101) had severe brain injury. The median time between the trauma and the interview was 300 days [Q1–Q3, 204–603]. For the purpose of the current study, we used these 97 issues as domains and chosed to consider that the 10 issues most frequently elicited by severe trauma survivors represented the most important domains for patients that should be considered in research16 (Table 1). These 10 patient-important domains were “I have pain,” “I have problems for walking,” “I can’t work like I used to,” “I am concerned about the consequences of the trauma on my future,” “I still have to perform regular medical tests or consultation,” “I get tired faster,” “I feel moral pain,” “I can no longer practice sports,” “I have memory problems,” “My appearance has been altered, I have scars.

TABLE 1. - Matrix Opposing the Measurement Tools Identified in the Literature Review and the Domains That Matter to Severe Trauma Survivors From the Qualitative Study
Domains Identified in the Qualitative Study—n (%) 10 Most Frequently Used Measurement Tools From the Literature Review
Number of Studies Using a Given Measurement Tools (%)
GOS or GOS (E)36 (31.0) Return to Work24 (20.7) EQ-5D22 (19.0) SF-3620 (17.2) SF-1215 (12.9) FIM12 (10.3) Work Status11 (9.5) PCL-C10 (8.6) Pain Numeric Scale8 (6.9) Surgical Complication7 (6.0)
I have pain (n = 170, 50.9%)
I have problems for walking (n = 115, 33.8%)
I can’t work like I used to (n = 75, 22.1%)
I am concerned about the consequences of the trauma on my future (n = 67, 19.7%)
I still have to perform regular medical tests or consultation (n = 63, 18.5%)
I feel moral pain (n = 56, 16.5%)
I get tired faster (n = 56, 16.5%)
I can no longer practice sports (n = 51, 15.0%)
I have memory problems (n = 48, 14.1%)
My appearance has been altered/ I have scars (n = 46, 13.5%)
Total number of construct(s) covered by the measurement tool 9 1 18 27 18 16 1 13 1 1
Mortality is not presented in the table because, in the qualitative study, we asked severe trauma survivors about their important health issues 6 months to 3 years after trauma and thus did not capture mortality as an issue that matter after trauma. Boxes shaded in gray indicate that the given domain (line) is captured by the given measurement tool (column).
Qualitative study was conducted among 340 participants. Return to work is defined as the possibility to work again (yes/no). Work status is defined as the need to change job, to adapt the job or to reduce the time at work.
FIM indicates functional independence measure; GOS; Glasgow Outcome Scale; GOS-E, Glasgow Outcome Scale—Extended; PCL-C, Posttraumatic Stress Disorder Checklist–Civilian;SF-36 or 12, Short Form Health Survey 36 or 12.

Matching Outcomes Identified During the Systematic Review and Domains From the Qualitative Study

For each health outcome retrieved in the systematic review, 1 author (AJ) looked for the original publication describing the measurement tool.31 Two authors (AJ, VTT) then matched each measurement tools with the domains identified in the qualitative study.16 For multiitem scales, where the domains were not evident, a third author blinded from other authors’ perspective performed the process once more. Discrepancies were resolved by discussion.

For example, the EQ-5D scale is used to capture the Health-Related Quality of Life and assesses pain, mobility (such as driving or walking), usual activities (such as working, family or leisure activities) anxiety, depression, and self-care (such as washing or dressing). Authors thus matched each components of the EQ-5D to the domains identified in the qualitative study such as “I’m limping,” “I can no longer drive,” “I have problems taking public transports,” “I have difficulty performing certain daily activities (recreational activity, house chores, shopping,...),” “I can no longer perform certain movements,” “I can no longer carry heavy loads,” “I can no longer practice sports,” “I am anxious in situations where the trauma could happen again,” “I feel moral pain” or “I have trouble getting dressed or washing myself or eating or going to the toilet.” Conversely, domains such “I don’t sleep well” would not be matched despite this latter domain being part of health-related quality of life and strongly correlated to the observed items of the EQ-5D.

The result from this step was a matrix opposing all measurement tools used in the literature to assess patients’ health after trauma to the domains that matter to severe trauma survivors (Table 1).16

Data Analysis

We evaluated the completeness of the collection of patient-important outcomes at study level and at field level. Completeness was reported as an array of 10 numbers, respectively, indicating the percentage of studies measuring at least 1 important domain among the 10 issues most frequently elicited by severe trauma survivors (ie, equivalent to “at least one PIO”); at least 2 domains; at least 3 domains, etc. Thus, the tenth value of the array indicates the percentage of studies measuring all domains considered as important (ie, the 10 most frequently elicited domains by severe trauma survivors in the qualitative study).

We evaluated the mismatch between domains assessed in the included studies and the domains elicited by participants during the qualitative study. We assessed which important domains, among the 10 issues most frequently elicited by severe trauma survivors in the qualitative study were among the 10 domains most frequently measured in studies.

We completed our analyses with a subgroup analysis opposing multicentric and monocentric studies. We then compared important domains frequency elicited by patients from the qualitative study with and without severe TBI (AIS head ≥3 and <3).

Analyses involved use of R v4.0.5 (http://www.R-project.org, the R foundation for statistical Computing, Vienna, Austria).

RESULTS

Step 1: Systematic Review of Outcomes Used in Reports Evaluating Long-term Outcomes After Trauma

We included 116 reports in the systematic review with 110 published studies and 6 protocols from clinical trials. About half of included reports (n = 55, 47%) were multicentric and 4% (n = 5) were randomized controlled trials (See Flowchart of selection of reports included in the review, Supplemental Material 4, see https://links.lww.com/AOSO/A174). Included published studies represented 105 961 patients with a median sample size of 225, interquartile range (IQR) 82 to 460. Lead author of all reports were mainly located in Europe (n = 56, 48%) or North America (n = 32, 28%). Reports characteristics are presented in Table 2 and Supplemental Material 5 (see https://links.lww.com/AOSO/A174).

TABLE 2. - General Characteristics of 116 Reports Included in the Systematic Review
Characteristic Value
Year of publication*, no (%)
 2014
 2015
 2016
 2017
 2018
 2019
15 (14)18 (16)11 (10)27 (25)21 (19)18 (16)
Location of the first author, no (%)
 Europe
 North America
 Asia & Oceania
 Africa
 Missing
56 (48)32 (28)25 (22)1 (1)2 (2)
Funding, no (%)
 Public
 Private
 Both
 None
 Unclear
82 (71)
3 (3)
2 (2)
4 (3)
25 (22)
Number of patients included*, median [Q1–Q3] 179 [82–460]
Multicentric study, n(%) 55 (47)
Randomized controlled trial, no (%) 5 (4)
Sex ratio, % men 77
Age* Mean† (SD): 43 (11)
Median ‡ [Q1–Q3]: 37 [33–41]
Injury severity score* Mean§ (SD): 22 (9)
Median∥ [Q1–Q3]: 22 [17–27]
Number of outcomes reported*, median [Q1–Q3] 3 [2–5]
Age, ISS, sex ratio was only available when reported for the whole study and not just within study groups.
*n = 110 published studies; †n = 61; ‡n = 19; §n = 28; ∥n = 17.

Included reports captured a total 403 outcomes, with a median number of measurement tools per reports of 3, IQR 2 to 5. The median time of primary outcome assessment in the included reports was 24 months, IQR 12 to 60. Data sources involved patient-reported data (80%, n = 324), administrative sources (9%, n = 35) or physician-reported data (8%, n = 31). For 13 reports, the data source was unclear (3%, n = 13). Using the taxonomy of Clarke and Williamson,29 most outcomes were classified as related to Function (26%, n = 106), Health-Related Quality of Life (23%, n = 93) and Mental Health (15%, n = 60) (Table 2).

These 403 outcomes were assessed using 154 unique measurement tools. Beside mortality (20.7%, n = 24 reports), the 3 measurement tools most frequently used to assess long-term health of trauma survivors in the 116 included reports were the Glasgow Outcome Scale or Glasgow Outcome Scale—Extended (31.0%, n = 36), homemade questions on patients’ ability to work (20.7%, n = 24) and the EQ-5D (5L or 3L) (19.0%, n = 22). These 3 measurement tools accounted for 20.3% (n = 82) of all measurement tools (n = 403) used in included reports. In contrast, among all unique measurement tools, 73.4% (n = 113) were used in only 1 reports. Table 1 presents the 10 most frequently used measurement tools and the most frequently elicited domains that matter to severe trauma survivors. Supplemental Material 6 (see https://links.lww.com/AOSO/A174) lists every measurement tool in the included reports.

Generic HRQOL tools, such as the EQ-5D32 or the SF3633 respectively covered 5 and 6 domains among the 10 most important. Specific trauma tools such as the POLytrauma Outcome Chart [POLO],34 the Quality Of LIfe in persons after traumatic BRain Injury [QOLIBRI]35 and the Trauma Outcome Profile36 covered 9 domains among the 10 most important while the Trauma Quality of Life measurement tools37 covered 8 of these 10 domains.

Step 2: Comparison Between the Outcomes Identified During the Systematic Review and Patients’ Perspective About Important Issues After Trauma

The matrix of 14 938 levels opposing the 154 measurement tools identified in the review to the 97 domains that matter to severe trauma survivors identified in the qualitative study was used to assess the completeness of the collection of patient-important domains (Fig. 1). Measurement tools used in the studies included in the review covered a median of 27 [2–36] domains. Among the included reports, 10.3% (n = 12) did not collect any patient-important domains while 69.8% (n = 81) collected at least 3 important domains and 26.7% (n = 31) collected at least 7 important domains (Fig. 1; Supplemental Material 7, see https://links.lww.com/AOSO/A174). Only 1 report collected at the same time all 10 domains considered as important38 and none of the 10 important domains were collected in more than two-thirds of the reports. Among patient-important domains, “I can’t work like I used to” was the most frequently collected in the reports included in the systematic review (n = 80, 69%). The mismatch between outcomes between domains assessed in the included reports and the domains elicited by participants during the qualitative study is presented in Figure 2.

F1
FIGURE 1.:
Presentation of the completeness of the collection of patient-important domains in the 116 reports assessing long-term outcomes after severe trauma. Patient-important domains are the 10 most frequently elicited domains in the qualitative study. They are each identified by a specific color. The idea is to show which domains were collected in each of the 116 reports included. Each horizontal line of the figure corresponds to 1 report included in the systematic review. The color bands show which of the 10 domains (labeled at the top) were reported for each report. Domains are sorted from the left (most frequently elicited, I have pain) to the right (10th most frequently elicited, My appearance have been altered). The 116 reports were sorted according to the total number of reported domains, in decreasing order with those reporting the most important number of domains at the top (n = 10/10 domains) to those reporting the less important number of domains below (n = 0/10 domains). The diagram on the left shows the distribution of the total number of reported domains across the 116 reports. The diagram at the bottom shows the proportion of reports that reported each domain.
F2
FIGURE 2.:
Mismatch between domains assessed in the included studies and domains that matter to patients identified in the qualitative study. Pink coloration indicate the 10 domains that matter to patients in the qualitative study (ie, patient-important outcomes, left column) and helps identifying their ranking among domains assessed in the included reports (right column).

Among the 10 domains most frequently elicited by severe trauma survivors during the qualitative study, only 5 “I can’t work like I used to,” “I can no longer practice sports,” “I have problem for walking,” “I have pain,”I feel moral pain,” and “I have problems walking” were among the 10 domains most frequently measured in included reports. The most important domain for patients in the qualitative study was “I have pain”; this domain was measured in 46% of included reports and ranked fourth of all measured domains. In contrast, the most frequently measured domain in the included reports was “I can’t work like I used to” (assessed 72% of included reports); it was elicited by 22% of patients in the qualitative study. Mismatch could be more important. For example, the domain “I still need to perform medical tests and consultations” was the fifth most important domain for patients, elicited by 19%; but was only the 52nd domain measured and was collected in 25% of reports.

The idea is to show which domains were collected in each of the 116 reports included. Each horizontal line of the figure corresponds to 1 report included in the systematic review. The color bands show which of the 10 domains (labeled at the top) were reported for each report. Domains are sorted from the left (most frequently elicited, I have pain) to the right (10th most frequently elicited, My appearance have been altered). The 116 reports were sorted according to the total number of reported domains, in decreasing order with those reporting the most important number of domains at the top (n = 10/10 domains) to those reporting the less important number of domains below (n = 0/10 domains). The diagram on the left shows the distribution of the total number of reported domains across the 116 reports. The diagram at the bottom shows the proportion of reports that reported each domain.

Subgroup Analysis

We highlight that the 10 domains most frequently elicited by participants’ with an AIS score for the head ≥3 and <3 were similar to those elicited by all the participants but for one (Supplemental Material 8, see https://links.lww.com/AOSO/A174). Comparing the completeness of the collection of patients-important domains according to both subgroups list of important domains found similar results (Table 3). Completeness in the collection of patients-important domains was greater among monocentric reports as compared to multicentric report as reported in Supplemental Material 7 (see https://links.lww.com/AOSO/A174).

TABLE 3. - Comparison of Number of Patient-important Outcome(s) Assessed by Included Studies According to the AIS Head of the Participants of the Qualitative Study
Based on Domains Elicited by All Patients Included in the Qualitative Study (n = 340 * ) Based on Domains Elicited by Patients With an AIS Head ≥ 3 (n = 125 * ) Based on Domains Elicited by Patients With an AIS Head < 3 (n = 214 * )
At least 1 104 (89.7) 104 (89.7) 104 (89.7)
At least 2 88 (75.9) 89 (76.7) 88 (75.9)
At least 3 81 (69.8) 79 (68.1) 81 (69.8)
At least 4 76 (65.5) 59 (50.9) 77 (66.4)
At least 5 54 (46.6) 50 (43.1) 56 (48.3)
At least 6 45 (38.8) 42 (36.2) 48 (41.4)
At least 7 31 (26.7) 27 (23.3) 40 (34.5)
At least 8 18 (15.5) 17 (14.7) 20 (17.2)
At least 9 11 (9.5) 12 (10.3) 12 (10.3)
At least 10 1 (0.9) 1 (0.9) 5 (4.3)
For each category, values are reported as n (%).
*AIS head was not reported for 1 participant.
AIS indicates abbreviated injury score.

DISCUSSION

Principal Findings

Studies assessing long-term health after severe trauma provide an incomplete picture of issues that matter to patients. Despite 89.7% of studies assessing at least one patient-important domain, only 46.6% measured at least 5 important domains. Moreover, 10.3% of studies did not capture any patient-important domain. Second, we found a striking mismatch between outcomes and domains measured in the literature and important domains for patients. For example, among the 10 most frequently measured domains in studies, only 3 are patients important.

Comparison With the Existing Literature

Until now, metrics used to evaluate the use of PIO in a given field have revolved around either assessing whether studies measure at least one PIO (which, in our case, would result in 89.7% of studies assessing a PIO)3,5,39 and studies assessing whether all outcomes from the corresponding COS are reported (which, in our case, would result in no (0%) study measuring all 10 patient-important domains).40 These extreme figures highlight that binary metrics do not adequately reflect the reality of clinical research. In complex conditions such as long-term health consequences of severe trauma, collecting only 1 PIO is not enough to capture patient-important issues and is unlikely to allow clinicians, patients, and decision makers to take informed decisions about policy and care.

Despite recommendations to involve patients in research, as of today, the definition of PIO mainly relies on physicians’ and researchers’ views.3,5,39 In this study, we used rich qualitative data obtained during semistructured interviews to define important issues after trauma. This allowed the identification of neglected domains, such as the fear of future consequences of the trauma on patients’ lives (fourth most important domain in the qualitative study, assessed by 28% of studies) or the burden of care after trauma (fifth most important domain in the qualitative study, assessed by 25% of studies).

The mismatch between outcomes captured in the literature and issues that matter to patients does not only concern research gaps. The most important domain for patients, “I have pain,” is only the 20th domain measured in studies that evaluate health after severe trauma and is evaluated in 61% of trials. Despite of numerous pharmacological and non- pharmacological interventions for pain, more than two-third patients still report chronic pain after severe trauma.41,42 A potential cause of this mismatch is that pain after trauma is complex to appraise, with polymorphic presentations such as nociceptive pain, postamputation pain, and complex regional pain syndrome.43 We believe that it is important to systematically assess pain in studies evaluating health after trauma. Multiple tools exist to capture these heterogeneous presentations of pain, ranging from simple tools (such as the pain numeric scale, the need for pain relief medications prescription or the EQ-5D) to more complex patient-reported outcomes (eg, the Trauma Outcome Profile, the Trauma Quality of Life measure or the Trauma Quality of Life Measurement tool).26

In 2020, Sakran et al proposed a consensus conference on Trauma Patient-Reported Outcomes measures.44 This work highlighted 10 PROMs to prioritize for assessing adult trauma patients’ health. When comparing these 10 PROMs to the measurement tools identified during the systematic review, we identified 4 were never captured in our systematic review (Patient-Reported Outcomes Measurement Information System, Quality of Life In Neurological Disorders, Alcohol Use Disorders Identification Test, Strengths and Difficulties Questionnaire). This underlines that important tools remain underused to date.

Strengths and Weaknesses of the Study

This study has several strengths. First, we used broad inclusion criteria, including both controlled and noncontrolled studies in the study period which provides an extensive description of outcomes currently used to assess long-term health after severe trauma. Second, in contrast to most studies assessing PIO, we avoided preconceptions in the definition of important issues after trauma by directly involving patients with qualitative methods. This led to the identification of unexpected important domains.

This study has several limitations. First, the matching process between the domains from the qualitative study and the measurement tools from the systematic review implies a certain degree of subjectivity. To ensure a replicable process, we made the methodological choice to assess multi item scales through their individual items and to assess whether each item covered (or not) the domains of the qualitative study. Therefore, such approach ignores that some scales use observed “traits” (in their items) to assess latent domains. For example, the EQ-5D is considered to assess mobility, pain, anxiety, usual activities, and self-care but not to assess the quality of sleep despite this latter domain being part of health-related quality of life and strongly correlated to the observed items of the EQ-5D. We also implied 3 experts in the matching process for multiitem scales and resolved disagreements by consensus

Second, we made the hypothesis that the results of the qualitative study, performed in France, are relevant to be applied in international studies, conducted in different health systems, countries, and contexts. Therefore, we may not adequately account for the differences in the severe trauma populations over the word especially regarding the proportion of penetrating injuries, trauma management, or potential financial implications for trauma survivors.

Third, the qualitative study included some participants (n = 68) with only 3 to 6 months recovery. However, none of the domains elicited by these patients were not specific to this population.

Fourth, we used broad inclusion criteria in our systematic review in order to provide a complete picture of outcomes and tools actually used after trauma. Because these inclusion criteria did not perfectly match those used in our qualitative study (ie, at least 1 AIS ≥ 3), we cannot exclude the risk that some patients included in the systematic review studies may be less severe than those from the qualitative study.

Fifth and last, when examining the measurement tools identified in the literature, we assessed whether each domain was assessed (or not). We did not appraise whether tools used a single or multiple items to capture the domain. For example, we considered that generic tools such as the POLO, the TOP, or the QOLIBRI, which included only 1 item on memory captured the domain “I have memory problems” and did not differentiate them from specific questionnaires focused on memory such as the GOAT.

Implications

We did not aim to judge the relevance of the measurement tools used in trauma research but rather highlight many issues considered important for patients are not captured in the current literature, after trauma. Our results highlight that most trauma survivors, including those with specific injuries such as TBI, seems to share common problems. By adding a limited number of patient-reported scales to all studies investigating the long-term consequences after trauma, in addition to more specific outcomes, we could drastically improve our ability to combine and compare study results. For example, specific trauma measurement tools such as the Trauma Quality of Life,37 the Trauma Outcome Profile or the POLytrauma Outcome Chart34 each cover at least 50% of the 97 domains identified in the qualitative study and at least 8 among the 10 most important. However, these measurement tools are long, complex and tedious for patients to complete and researchers must take feasibility into consideration. Conversely, generic measurement tools such as EQ-5D, SF-36, or SF-12 are easier to collect and to compare across the medical literature but may overlook almost half of the trauma survivors’ important domains. One solution for researchers deciding to use generics tools while wishing to maintain an important completeness is to combine them with simple and trauma specific measurement tools such as surgical complications, posttraumatic stress disorder scales. For example, simply including the SF-12, the Posttraumatic Stress Disorder Checklist for Civilian (PCL-C), and 2 questions about surgical complications and appearance would allow to cover all 10 important domains considered in this study. Second, our study introduces new concepts to appraise the use of PIO in the clinical literature, moving from limited binary metrics to the conceptualization of what matters to patients, in a given clinical situation, as a whole. Indeed, in reviews assessing whether studies at least 1 PIO, measuring at least “Pain” or at least “Administrative burden” provides a similar result. In our approach, each PIO become unique and complementary to others, which justify an appraisal of the completeness of a set of PIO. At the field level, this allows the identification of mismatch, with some important outcomes rarely or never assessed in the literature.

Finally, our study can be seen as the foundation for a COS for long-term outcomes after severe trauma. By combining the findings from a large qualitative study with a systematic review of all outcomes measured in the field, we provide researchers with useful data to prioritize and standardize outcomes to be collected in studies, with reason, the potential to reduce waste in research.45–47

CONCLUSION

By combining the results of a qualitative study conducted in France and a systematic review, we introduce the notion of completeness of the collection of patient-important domains in research. We show that the field of long-term consequences after trauma is not complete, with an important mismatch between what matters to patients and outcomes used in the literature. The capture of these unmeasured outcomes could change physicians’ perceptions of the interventions they prescribe or of the health status of their patients. Beyond the clinical message, this study thus advocates for patients’ involvement in the definition of outcomes to be used in research.

ACKNOWLEDGMENTS

We thank Elise Diard for her precious help in drafting figure and Caroline Barnes for English proofreading. We also thank Pauline Glasman, Nicolas Gatulle, and Cyril Quemeneur for their help in the matrix management.

REFERENCES

1. Meinert C. Clinical Trials Dictionary: Terminology and Usage Recommendations, 2nd Edition. Wiley; 2012.
2. Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallel group randomized trials. BMC Med Res Methodol. 2001;7:1–7.
3. Gandhi GY, Murad MH, Fujiyoshi A, et al. Patient-important outcomes in registered diabetes trials. JAMA. 2008;299:2543–2549.
4. Rahimi K, Malhotra A, Banning AP, et al. Outcome selection and role of patient reported outcomes in contemporary cardiovascular trials: systematic review. BMJ. 2010;341:c5707.
5. Gaudry S, Messika J, Ricard JD, et al. Patient-important outcomes in randomized controlled trials in critically ill patients: a systematic review. Ann Intensive Care. 2017;7:28.
6. Hay SI, Abajobir AA, Abate KH, et al. Global, regional, and national disability-adjusted life-years (DALYs) for 333 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. The Lancet. 2017;390:1260–1344.
7. Harmsen AM, Giannakopoulos GF, Moerbeek PR, et al. The influence of prehospital time on trauma patients outcome: a systematic review. Injury. 2015;46:602–609.
8. Celso B, Tepas J, Langland-Orban B, et al. A systematic review and meta-analysis comparing outcome of severely injured patients treated in trauma centers following the establishment of trauma systems. J Trauma. 2006;60:371–8; discussion 378.
9. Raux M, Le Manach Y, Gauss T, et al.; TRAUMABASE Group. Comparison of the prognostic significance of initial blood lactate and base deficit in trauma patients. Anesthesiology. 2017;126:522–533.
10. Harrois A, Soyer B, Gauss T, et al.; Traumabase® Group. Prevalence and risk factors for acute kidney injury among trauma patients: a multicenter cohort study. Crit Care. 2018;22:344.
11. Gauss T, Ageron FX, Devaud ML, et al.; French Trauma Research Initiative. Association of prehospital time to in-hospital trauma mortality in a physician-staffed emergency medicine system. JAMA Surg. 2019;154:1117–1124.
12. Holcomb JB, Tilley BC, Baraniuk S, et al.; PROPPR Study Group. Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: the PROPPR randomized clinical trial. JAMA. 2015;313:471–482.
13. Pusateri AE, Moore EE, Moore HB, et al. Association of prehospital plasma transfusion with survival in trauma patients with hemorrhagic shock when transport times are longer than 20 minutes: a post hoc analysis of the PAMPer and COMBAT clinical trials. JAMA Surg. 2020;155:e195085.
14. Davidson GH, Hamlat CA, Rivara FP, et al. Long-term survival of adult trauma patients. JAMA. 2011;305:1001–1007.
15. Sierink JC, Treskes K, Edwards MJ, et al.; REACT-2 study group. Immediate total-body CT scanning versus conventional imaging and selective CT scanning in patients with severe trauma (REACT-2): a randomised controlled trial. Lancet. 2016;388:673–683.
16. James A, Tran VT, Gauss T, et al. Important issues to severe trauma survivors: a qualitative study. Ann Surg. 2020;1:189–195.
17. Brohi K, Schreiber M. The new survivors and a new era for trauma research. PLoS Med. 2017;14:e1002354.
18. Gargon E, Williamson PR, Altman DG, et al. The COMET initiative database: progress and activities update (2014). Trials. 2015;16:515.
19. Matvienko-Sikar K, Avery K, Blazeby J, et al. Use of core outcome sets was low in clinical trials published in major medical journals. J Clin Epidemiol 2022;142:19–28.
20. Young B, Bagley H. Including patients in core outcome set development: issues to consider based on three workshops with around 100 international delegates. Res Involv Engagem. 2016;2:25.
21. Kirwan JR, Minnock P, Adebajo A, et al. Patient perspective: fatigue as a recommended patient centered outcome measure in rheumatoid arthritis. J Rheumatol.:4.
22. Sinha IP, Gallagher R, Williamson PR, et al. Development of a core outcome set for clinical trials in childhood asthma: a survey of clinicians, parents, and young people. Trials. 2012;13:103.
23. Davis K, Gorst SL, Harman N, et al. Choosing important health outcomes for comparative effectiveness research: An updated systematic review and involvement of low and middle income countries. PLoS One. 2018;13:e0190695.
24. Zarin DA, Tse T, Williams RJ, et al. The ClinicalTrials.gov results database–update and key issues. N Engl J Med. 2011;364:852–860.
25. Chan AW, Tetzlaff JM, Altman DG, et al. SPIRIT 2013 statement: defining standard protocol items for clinical trials. Ann Intern Med. 2013;158:200–207.
26. Hoffman K, Cole E, Playford ED, et al. Health outcome after major trauma: what are we measuring? PLoS One. 2014;9:e103082.
27. Pape HC, Lefering R, Butcher N, et al. The definition of polytrauma revisited: An international consensus process and proposal of the new ‘Berlin definition’. J Trauma Acute Care Surg. 2014;77:780–786.
28. Centers for Disease Control and Prevention. Traumatic Brain Injury In the United States: Epidemiology and Rehabilitation.https://www.cdc.gov/traumaticbraininjury/pdf/TBI_Report_to_Congress_Epi_and_Rehab-a.pdf. Published 2016.
29. Dodd S, Clarke M, Becker L, et al. A taxonomy has been developed for outcomes in medical research to help improve knowledge discovery. J Clin Epidemiol. 2018;96:84–92.
30. Porter ME, Larsson S, Lee TH. Standardizing patient outcomes measurement. N Engl J Med. 2016;374:504–506.
31. Lavrakas P. Encyclopedia of Survey Research Methods. 2008. Available at: https://methods.sagepub.com/reference/encyclopedia-of-survey-research-methods/n91.xml.
32. Herdman M, Gudex C, Lloyd A, et al. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011;20:1727–1736.
33. Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30:473–483.
34. The POLO Chart Study Group, Lefering R, Tecic T, et al. Quality of life after multiple trauma: validation and population norm of the Polytrauma Outcome (POLO) chart. Eur J Trauma Emerg Surg. 2012;38:403–415.
35. von Steinbuechel N, Petersen C, Bullinger M, The QOLIBRI Group. Assessment of health-related quality of life in persons after traumatic brain injury — development of the Qolibri, a specific measure. Acta Neurochir (Wien). 2005;93:43–49.
36. Attenberger C, Amsler F, Gross T. Clinical evaluation of the Trauma Outcome Profile (TOP) in the longer-term follow-up of polytrauma patients. Injury. 2012;43:1566–1574.
37. Wanner JP, deRoon-Cassini T, Kodadek L, et al. Development of a trauma-specific quality-of-life measurement. J Trauma Acute Care Surg. 2015;79:275–281.
38. Kaske S, Lefering R, Trentzsch H, et al. Quality of life two years after severe trauma: a single-centre evaluation. Injury. 2014;45 Suppl 3:S100–S105.
39. Ameur H, Ravaud P, Fayard F, et al. Systematic reviews of therapeutic interventions frequently consider patient-important outcomes. J Clin Epidemiol. 2017;84:70–77.
40. Hughes KL, Kirkham JJ, Clarke M, et al. Assessing the impact of a research funder’s recommendation to consider core outcome sets. PLoS One. 2019;14:e0222418.
41. Nampiaparampil DE. Prevalence of chronic pain after traumatic brain injury: a systematic review. JAMA. 2008;300:711–719.
42. Rivara FP, Mackenzie EJ, Jurkovich GJ, et al. Prevalence of pain in patients 1 year after major trauma. Arch Surg. 2008;143:282–7; discussion 288.
43. Beard DJ, Aldington D. Chronic pain after trauma. Trauma. 2012;14:57–66.
44. Sakran JV, Ezzeddine H, Schwab CW, et al.; Patient Reported Outcome Consortium. Proceedings from the consensus conference on trauma patient-reported outcome measures. J Am Coll Surg. 2020;230:819–835.
45. Chalmers I, Glasziou P. Avoidable waste in the production and reporting of research evidence. Lancet. 2009;374:86–89.
46. Macleod MR, Michie S, Roberts I, et al. Biomedical research: increasing value, reducing waste. Lancet. 2014;383:101–104.
47. Ioannidis JP. Why most published research findings are false. PLoS Med. 2005;2:e124.
Keywords:

patient-important outcomes; multiple trauma; patient outcome assessment; long-term care

Supplemental Digital Content

Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.