Posttraumatic stress symptoms indicative of PTSD were associated with more problems on almost all domains of functional outcome and a considerable decrease of HRQoL in both non-hospitalized and hospitalized injury patients two years post-injury.
Previous studies on PTSD and HRQoL were conducted in clinical patient populations and were therefore restricted to accidents and injuries at the higher end of the severity spectrum [20
]. This study was not restricted to particular injury subgroups, such as adolescent victims or victims with severe injuries [4
]. The high variety in injuries included in this study and the relatively large sample size allowed examination of the association of a number of injury characteristics and posttraumatic stress symptoms indicative of PTSD.
We found that injury patients with posttraumatic stress symptoms indicative of PTSD reported significantly more problems on all EQ-5D and almost all HUI3 health domains. A study that investigated HRQoL with EQ-5D among patients with PTSD following cardiac arrest reported similar findings [39
]. Among adolescent victims PTSD was associated with impairments in Role/Social Behavioral, Role/Social Physical, Bodily Pain, General Behavior, Mental Health, and General Health Perceptions subscales of the 87-item Child Health Questionnaire [22
]. The resulting EQ-5D and HUI3 utility scores of injury patients with PTSD found in the current study are approximately in the range of the utility scores that Holbrook et al. derived with the multi-attribute utility instrument Quality of Well-being scale (QWB) (0.58 - 0.62) [20
]. Although the HUI3 instrument yielded significantly lower health utility scores compared to the EQ-5D, which accords with results of other studies [40
], both HUI3 and EQ-5D showed that PTSD was associated with a mean utility loss of 0.17 - 0.25. This concurs with the utility loss of anxiety disorders social phobia, generalized anxiety disorder and agoraphobia [43
It should be noted that Holbrook et al. focused on injury patients admitted to a trauma centre with a length of stay of more than 24 hours and patients injured due to unintentional and intentional injury, whereas the current study included all admitted injury patients to general and university hospitals who were injured due to unintentional injury. Moreover, Holbrook et al. used an IES-score greater than 24 to identify patients with PTSD, whereas in the current study a cut off of 35 was used. Evidence from studies on this matter suggests that to avoid overestimation of the number of cases with PTSD, an IES-score of greater than 35 is more appropriate [26
]. Using the DSM-IV as the diagnostic criteria for PTSD, a cut-off score of 35 produced sensitivity of .89, specificity of .94 [26
]. With a cut-off point of 24, the sensitivity is 0.91 and the specificity 0.46 [45
]. To avoid over diagnosing of PTSD in a comprehensive population with a relative low PTSD prevalence, it is important to use a high IES cut-off score that incurs a high specificity.
Nonetheless, an important shortcoming of this study was that existence of PTSD symptoms was measured with the IES rather than Clinician-Administered PTSD Scale for DSM-IV (CAPS). The IES is a self-report questionnaire that measures only two of the three main PTSD symptoms, namely intrusion and avoidance. It is not a diagnostic tool, i.e., it is not designed to diagnose mental disorders according to the DSM-IV (the fourth edition of the diagnostic and statistical manual for psychiatric disorders). Consequently, cases that in the current study were identified as having PTSD symptoms might not meet the DSM-IV criteria of clinical PTSD, and inversely. Due to differences in assessment of PTSD symptoms it is difficult to compare the results found in this study to previous studies on PTSD and health-related quality of life (HRQoL).
Both hospitalized and non-hospitalized injury patients with symptoms indicative of PTSD at 24 months post-injury reported a decrease in health status after 9 months, which may indicate that the sample is starting a deterioration process. On the other hand, patients might have overestimated their 9-month health status (and possibly also their 5-month health status), because their frame of reference has changed as a result of a temporary decrease in health status after the injury (response shift) [46
]. However, without information on PTSD status at previous measure points, the reasons for the reductions in HRQoL at 24-months post-injury can only be speculated about.
Functional consequences of injury, both temporary and permanent, show large variations dependent on the injury location and injury type. In the current study we used the European injury classification EUROCOST [48
]. This classification is compatible with the International Statistical Classification of Diseases, Injuries and Causes of Death - Ninth revision (ICD-9) classification system an consists of 39 injury groups that are homogeneous in terms of healthcare use, disability, as well as treatment and prognosis. In terms of anatomical classification the EUROCOST classification is simple compared to the ICD, which provides very detailed information on injury diagnoses by location and type of injury.
A second limitation of this study was the low response rate of the follow-up questionnaires [23
]. The 24-month follow-up questionnaire, which included the IES, was send only to those patients who responded to the preceding three follow-up questionnaires send at 21/2, 5 and 9 months. This meant that only 21% of the patients of the initial sample selected for the follow-up study filled in the 24-month follow-up questionnaire. However, the data were adjusted for non-response and possible response bias, because the PTSD prevalence rates were calculated using data that were weighted with respect to the original sample size and composition by inverse probability weighting. For some aspects, such as the severity of sustained injuries, the adjustments of non-response could be improved, since injury severity scores were not available.
Evidence suggested that patients with very severe health problems are less likely to respond to a survey [49
]. Differential underreporting by level of severity cannot be excluded, since we found a larger proportion of hospitalized patients among those with PTSD at 2 years post-injury. This could partly be caused by missing a larger share of the more severely injured hospitalized patients among those without PTSD (e.g. comatose patients). This may have led to a slight overestimation of the utility losses due to PTSD. However, severely injured patients are only a minor part of the total sample and PTSD remained significantly associated to adverse HRQoL, even after adjustment for confounders including hospitalization status.
In the current study PTSD is measured at 24 months follow-up only. A longitudinal study on PTSD and HRQoL among injury patients might elucidate any causal relationship between PTSD and subsequent reduced HRQoL. Furthermore, the influence of earlier HRQoL on PTSD remains to be investigated.