1
How does changes in symptom severity
index relate to patients’ global
impression of change?
Sandra Vestlundand Linn Nykvist Tutors: Anders Wänman and Aurelia Ilgunas
Examensarbete, 30hp
Faculty of Medicine, Department of Odontology, Umeå University Umeå, Sweden 2020
Abstract: 253 Text: 4003
2
ABSTRACT
Background: The treatment outcome of temporomandibular disorders (TMD) can be
assessed with Symptom Severity Index (SSI). In addition to this self-report measure, Patients’ Global Impression of Change (PGIC) provides information about the overall impression of change. The use of SSI and PGIC when assessing treatment outcome in the field of TMD, and their correlation, has not previously been reviewed.
Aim: The aims of this study were to investigate the correlation between the scales SSI and
PGIC, to gain knowledge about which factors that contribute to the impression of change and to identify the diagnoses that have the most improvement.
Methods: Data from 193 patients was analyzed. The percentage change in symptoms
between baseline and follow up was compared to PGIC.
The patients were divided into three groups based on main diagnoses, for further analysis. PGIC categories “much improved” and “very much improved” were set as successful treatment outcome. Different factors impact on PGIC were analyzed.
Results: Majority of the patients had a successful treatment outcome. A higher mean
percentage change correlated with a better PGIC value. Patients with articular related diagnoses needed a greater reduction in SSI to rate the impression of change as improved. According to the regression analysis, diagnosis group, age, number of treatment visits and SSI-baseline were associated with treatment outcome.
Conclusions: The present study supports the use of both SSI and PGIC for assessment of
3
BACKGROUND
Temporomandibular disorders (TMD) is a heterogeneous group of functional disturbances involving the jaw muscles, the temporomandibular joint (TMJ) and associated structures (Hoffmann RG et al. 2011). Patients with TMD often experience pain localized to the TMJ or jaw muscles, limited jaw motions and TMJ-sounds like popping or grating sounds (Scrivani SJ et al. 2008). The onset of TMD may be acute and the symptoms mild and transient. In other cases, the condition becomes chronic followed by persistent pain and severe jaw dysfunction (Hoffmann RG et al. 2011).
Comorbid conditions are common among patients with TMD and the chronic conditions may include symptoms commonly found in patients with other chronic pain syndromes in other parts of the body, such as chronic tension-type headache or migraine, lower back pain, and fibromyalgia (Scrivani SJ et al. 2008, List T et al. 2017). TMD, coexisting pain and comorbid conditions may share a common etiology (Furquim BD et al. 2015). Even though the
pathophysiology is not clearly understood, a generally accepted mechanism behind the more chronic conditions is central sensitization, which causes amplified pain and perception of pain (Nguyen TT et al. 2019).
The etiology is complex, and a biopsychosocial approach is widely acknowledged to explain the cause of TMD (Edwards RR et al. 2016). The different conditions are considered the outcome of physiological, psychological and social factors with comorbid conditions overlapping each other. These factors may predispose, initiate or perpetuate TMD and contribute to the experience of pain and dysfunction (List T et al. 2017). Due to the complex and heterogeneous nature of TMD, the disorders can manifest in various ways, with several symptoms and therefore require different treatment approaches. These factors may
contribute to varying treatment outcomes.
4 NRS is commonly used when assessing intensity of symptoms. The scale ranges from 0-10 (0= no pain), (10= worst possible pain). It is simple and easy for patients to understand. Farrar and colleagues demonstrated that the NRS for pain is a valid indicator of the patient’s perception of improvement in pain intensity, in a study with Duloxetine treatment of chronic pain disorders (Farrar JT et al. 2010). Pain occurring daily or monthly will have different impact on the experience of pain. Measure of pain frequency is an important aspect of the experience of pain and has validity although it has not received much attention in pain research (Dworkin RH et al. 2005). Positive changes in the self-assessed pain measurements indicate improvement. A statistically significant improvement does however not necessarily mean a perceived, meaningful improvement for the patient (Salaffi F et al. 2004). The goal of any treatment should thus be the subjective impression of improvement.
Another measure for evaluating the outcome of treatment is the scale Patients’ Global Impression of Change (PGIC). PGIC is commonly used to evaluate the overall improvement or change in symptoms following treatment and is an important indicator of significant change in treatments of chronic pain (Scott W et al. 2015). The patient is asked to evaluate the impression of change in symptoms, since the beginning of treatment, on a 7-point scale ranging from -3 (very much worsened) to +3 (very much improved). PGIC can be used for the overall improvement or for single dimensions. Evaluating the overall improvement provides little information about which symptom that has improved and thereby contributed to the impression of change. Since patients with TMD may experience numerous symptoms the patient is asked to rate the PGIC for several dimensions. The study by Gagnon and colleagues support the use of an expanded model of PGIC (Gagnon CM et al. 2018).
The literature contains examples of both underestimation and overestimation of recalled baseline pain, although the latter is more common. It has been suggested that present pain easily affects the rating of recalled pain (Salovey P et al. 1993). This indicates that an
individual with a substantial improvement but with remaining pain at follow up may rate the change as minimal or non-existent, whereas an individual with a small raw change but no present symptoms may rate the change as “very much improved”.
5 The primary aim of this study was to investigate the correlation between the clinical rating scales Symptom severity index (SSI) and Patients’ Global Impression of Change (PGIC), by comparing patients’ rated pre-post treatment change in symptoms with their global
perception of change after treatment.
A secondary aim was to identify the diagnoses that have the most improvement and which factors that have the largest impact on patients’ impression of change. We hypothesized that less improvement is seen in patients with diagnoses related to the TMJ and that women experience less improvement than men (Wahlund K, Larsson B 2017). We also predicted a correspondence between SSI and PGIC.
METHODS
Study design and participants
This study is an observational study for quality assurance.Data collected from digital dental records of 342 patients referred and consecutively examined at the Department of Clinical Oral Physiology, Public Dental Health service, in Västerbotten County, Sweden, were analyzed.
All subjects had been examined and treated by the same dental resident in Clinical Oral Physiology during 2014-2019. The patients received treatment for different diagnoses of TMD, such as local, regional and widespread myalgia, arthralgia, arthrosis and disc
displacement with or without reduction. See Tabel 1 for all included diagnoses. The treatment varied depending on diagnosis, and in many cases, several treatment methods were
combined. Treatment forms used were information, jaw exercises, stretching, sensory stimulation (massage and heat), pharmacological treatment (NSAID, glucocorticosteroids and anesthetics), different kinds of occlusal splints, occlusal correction and/or
temporomandibular joint surgery.
Measures
Patients referred to the department of Clinical Oral Physiology at the Department of Clinical Oral Physiology, Public Dental Health service, in Västerbotten County, Sweden, were
assessed at baseline using the Research Diagnostic Criteria for TMD (RDC/TMD), Diagnostic Criteria for TMD (DC/TMD) and International Classification of Diseases (ICD 10) (List T, Højland Jensen R 2017). In addition to this evaluation SSI and PGIC, both related to the individual experience of the symptoms, were filled out at baseline and follow-up.
6 a higher severity level of symptoms. After treatment, at the follow-up, a new SSI was counted to find out if the symptoms had changed for the better or the worse following treatment. At the follow-up in addition to SSI a second scale, PGIC, was used. The patient was asked to rate the subjective overall impression of change, since beginning of treatment, on a 7-point scale (Figure A2).
Inclusion criteria for analysis
•
Symptom severity index > 0 at baseline.•
Symptom severity index at follow up.•
Patients’ Global Impression of Change at follow-up.Subjects with a SSI = 0 at baseline were excluded. These subjects either had incomplete data or lacked subjective symptoms at start. Subjects with a missing SSI value at follow-up or a missing value in PGIC were also excluded from data analyses.
Data analysis
IBM SPSS version 25 was used for data analyses. Descriptive statistics were used for percentages, means and standard deviations and frequencies were used for assessing the number of individuals in different categories.
Change between SSI baseline and SSI follow-up were presented both in raw change and in percentage change (raw change/SSI baseline x 100). Raw change was calculated by
subtracting SSI follow up from SSI baseline. These two ways of presenting change can vary depending on the severity of the symptoms. The association between the scales was analyzed to evaluate if one of them was enough for the analyses.
The bivariate association between percentage change and raw change in SSI was analyzed using the Spearman rank correlation coefficient.
Percentage change and PGIC were also compared with a Spearman rank correlation coefficient to analyze the association.
7 pain and jaw dysfunction associated with WAD. These diagnoses were combined into the third mixed group (n=17). Because of the lack of consistency this heterogeneous group was only part of analyses regarding the total data.
Differences in percentage change in SSI between diagnoses related to TMJ and diagnoses related to jaw muscles were estimated using Mann-Whitney test on two independent samples.
The data from PGIC was dichotomized (PGIC <2 and PGIC 2-3) for further analysis. The items “much improved” (2) and “very much improved” (3) on the PGIC-scale were set as criterion for a good treatment result, since these categories can be considered “treatment success” or “satisfactory improvement” (Dworkin RH et al. 2008). A demarcation was also necessary for statistical analyses.Finally, different factors influence on patients’ PGIC ratings were analyzed in a multivariate logistic regression using the new variable of PGIC as
dependent variable. The following covariates were added to the equation; Sex, age, diagnosis-group (TMJ-related/muscular-related/other), number of treatment visits and SSI-baseline. The significance level was set at <0.05.
To get an idea of how many of the patients that experienced any kind of improvement, i.e. minimally improved included, the total percentage of individuals with a positive rating on the PGIC scale was calculated.
The articles referred to in this study were searched by free-terms and MESH-terms in the PubMed database. No language restriction was used. Literature was also found from reference lists of articles found on PubMed and by Google searching. Terms used for the PubMed search were: “Pain measurements and treatment outcome and PGIC”, “PGIC and temporomandibular disorders”, “Important changes and PGIC”, “Temporomandibular joint disorders and risk factors and comorbidity”,“Temporomandibular joint disorders and chronic pain”, “Intensity and frequency in temporomandibular joint disorders”,
“Development of chronic pain”, “Important changes in the numerical rating scale”. Articles selected for evaluation had similar self-report methods or were of interest for the field of pain and dysfunction.
Ethical reflection
8 This is supported by The Patient Data Act SFS 2008:355 Ch 7, 4§ and 5§ since it concerns quality assurance. A subjective disadvantage is the violation of personal integrity when collecting data without notifying the subjects, but the overall benefit was estimated to be greater than the disadvantage for the individual patient.
The results provide a greater understanding for different factors influence on patients’ impression of illness and may be beneficial for improvement of future treatment outcome. The Ethics committee at the Department of Odontology, Umea University, approved the study as a student project.
RESULTS
The flowchart presents the 193 patients included in data analysis, and reasons for exclusion (Figure 1). The sex distribution in the initial data and the final data, after exclusion, were similar with the women to men ratio 2,7 in the initial data and 2,6 in the final data. The majority of the patients with incomplete data had declined or canceled treatment and some of them had been referred to other instances. Ages of the examined patients ranged from 11 to 88 years with a mean age of 35 years. Diagnoses in the final data is presented in Table 1, together with mean SSI at baseline and follow up and percentage change for the different diagnosis groups. A general improvement after treatment was observed with a decreasing mean SSI from baseline to follow-up (Table 1). Raw change and percentage change had a close association (Spearman correlation=0.57, significance level 0.01) and therefore percentage change was used in the analyses.
According to PGIC, majority of the patients rated their change as “much improved” or “very much improved” (n=130, 67%) and more women (72%) than men (56%) experienced this kind of improvement, comparable to successful treatment. If the category “minimally
improved” is included in the group of improved patients, almost 90 % are considered to have a positive treatment outcome.
A correlation between PGIC and the percentage change in SSI was supported by the correlation test (Spearman r=0,58, p=0,00 significant at the 0,01 level).A comparison between PGIC and percentage change in SSI for the total data is presented in graphical format in Figure 2. The distribution of the percentage change in SSI (min, max, mean and SD) corresponding to each PGIC-category for the total data is presented in table 2.
9 The Mann-Whitney test showed a statistical difference in percentage change (p=0.001) between the two diagnosis groups (Arthrogenic-diagnosis/Muscular-diagnosis).
When data from PGIC was dichotomized, 130 (67%) patients had rated their impression of change to “much improved” and “very much improved and 63 (33%) patients had rated their impression of change as “much worse” to “minimally improved” (no subjects in “very much worse and “minimally worse”). Only one person rated the change as “much worse” on the PGIC-scale, a woman with allodynia.
According to the multivariate regression analysis, diagnosis group, age, number of treatment visits and SSI-baseline were associated with treatment outcome (Table A1).
DISCUSSION
In this study, we examined the relationship between patients’ rated change in SSI and their impression of change after receiving treatment. The results showed that patients’ impression of change essentially match changes in symptom intensity and frequency, as we
predicted. This implies that both scales are valid tools when assessing treatment outcome in the treatment of TMD and that the combination of them is useful in the clinical setting.
This study was a part of quality assurance and the results indicate that the level of the care provided maintains a high standard. The majority of the patients experienced an
improvement following treatment, which proves that the treatment was successful.
Change in SSI was primarily presented with percentage change because of the close
association between percentage change and raw change. The use of percentage change rather than raw change as primary measure of actual change is supported by other studies (Emshoff R et al. 2010). Farrar and colleagues have recommended the use of percent change in studies with great variability in baseline pain since raw change is less consistent when compared with PGIC (Farrar JT et al. 2001). The percentage change scores presented a wide distribution in all the PGIC-categories. However, a higher mean percentage change correlated with better impression of change.
10 indicates almost total symptom relief which means they’ve reached the largest improvement possible. Only one person in this category had a 0% change.
The majority of the patients (67%) had a successful treatment outcome and had rated the change as “much improved” or better, which is considered a clinically important
improvement and was therefore set as a criterion(Farrar JT et al. 2000).
Farrar and colleagues have compared 10 studies on chronic pain and concluded that a 30% reduction in NRS is best associated with the PGIC category “much improved” (Farrar JT et al. 2000). Our results indicate that a greater percent change is needed when including
frequency of symptoms in the equation, since the SSI-scale contains both the NRS and frequency. For the articular related diagnoses an 80% mean change was required for rating the impression of change to “much improved” compared to muscular related diagnoses where a 60% mean change seemed to be enough. However, there was a great distribution between different individuals, especially in the muscle group, which shows the difficulty in using self-assessment scales. We can only speculate why a higher reduction in percent change is required when including frequency of symptoms. The construct of the scales may have an impact and it becomes more complex when another dimension of the symptom is evaluated. In this study we did not have access to information regarding intensity and frequency separately. Analyzing these variables separately could explain their individual impact on PGIC ratings. However, other studies have demonstrated that intensity is the most important factor contributing to an improvement in symptoms regardless of underlying disease,
treatment modality, age or gender (Farrar JT et al. 2001, Perrot S, Lantéri-Minet M 2019).
Our results suggest that patients with articular diagnoses require a greater percentage change for each step on the PGIC-scale compared to muscular diagnoses. All individuals with
articular related diagnoses had a reduction in SSI, although some of them had rated their impression of change as “unchanged”.
It is important to have in mind that TMD is multidimensional, including a great
11 that the main target for treatment may impact the domain of focus when assessing the
impression of change. (Perrot S, Lantéri-Minet M et al. 2019)
The impact of function on impression of change can also explain why patients with TMJ-related diagnoses seem to need a greater reduction in symptoms, to rate the impression of change with a higher value on the PGIC-scale. Dysfunction such as TMJ-sounds, catching or locking of the jaw is present in many of the TMJ-related diagnoses; hence, patients with any form of functional disturbances might need a larger change to experience a meaningful improvement.
The similar sex distribution in the initial data and final data, including data from follow up, indicates that sex does not impact weather the patient agrees to treatment or not.
Contrary to our hypothesis, more women experienced an improvement and rated their change as “much improved” or “very much improved”. Similar findings have been published in an article by Shinal and Fillingim (Shinal RM et al. 2007). They suggest that women have greater response to treatment and in particular when treatment is multimodal, which is the case in this study.
Identified shortcomings in this study are the great variability in diagnoses and treatments which complicated the analyses. All diagnoses with muscular origin were related to pain while the TMJ group was more heterogeneous with both pain and functional disturbances as popping, grating and locking of the jaw. Some of the diagnoses related to TMJ were
associated with an underlying disease, which also influences the treatment response. The treatment varied between the patients and treatments were often combined which made it difficult to compare them to each other. As a consequence, we did not analyze the various treatments impact on treatment outcome, and it should be subject to further analyses in future studies.
There was a substantial difference between the percentage change scores and raw change scores for all the PGIC groups. Depending on baseline scores, both small and great improvements in SSI can result in a 100% change whereas raw change shows the actual numerical change. For example, two individuals with very different raw change can have the same percentage change in symptoms. However, as mentioned earlier, statistical analysis still shows a correlation between the two of them.
The data we had access to did not contain any information about the duration of the
12 pain has poorer treatment outcome and that psychosocial factors can perpetuate the
symptoms and possibly affect the treatment outcomenegatively (Huttunen J et al. 2019). These factors could influence the ratings.
Conclusively, the present study supports the use of both SSI and PGIC for assessment of the treatment outcome when treating TMD. If there is a substantial reduction in symptoms on the SSI-scale and the patient rates the symptoms as much improved the dentist can be positive there is a subjective meaningful treatment outcome. Disagreements in SSI and PGIC ratings however brings important information.That is if a patient has rated the symptoms equivalent to a 100% improvement but experiences the change as non-improved or even worsened. Disagreements can indicate that an inaccurate treatment has been carried out or that the patient is incorrectly diagnosed. It also indicates that further examination is required to obtain information about underlying issues. If this is the case, it’s important to reassure that the patient understands the scales correctly to avoid misunderstandings. On the other hand, in reverse cases, if the SSI change is small or even decreasing but the patient’s impression of change is positive the dentist can count the treatment as successful and helping, even though it is more of a placebo effect.
Percentage change in combination with PGIC gives a clear image of the treatment effect and the relationship between the patient´s impression of change and rated symptoms. The percentage change could be noted in the patient’s medical record at the end evaluation after treatment.
ACKNOWLEDGMENTS
13
REFERENCES
Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA, Jensen MP, Katz NP, et al. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain. 2005; 113: 9–19.
Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, et al. Interpreting the Clinical Importance of Treatment Outcomes in Chronic Pain Clinical Trials: IMMPACT Recommendations. J Pain. 2008; 9: 105–121.
Edwards RR, Dworkin RH, Sullivan MD, Turk DC, Wasan AD. The Role of Psychosocial Processes in the Development and Maintenance of Chronic Pain. J Pain 2016;17: T70–92.
Emshoff R, Emshoff I, Bertram S. Estimation of clinically important change for visual analog scales measuring chronic temporomandibular disorder pain. J Orofac Pain. 2010; 24: 262– 269.
Farrar JT, Portenoy RK, Berlin JA, Kinman JL, Strom BL. Defining the clinically important difference in pain outcome measures. Pain. 2000; 88: 287–294.
Farrar JT, Young JP, LaMoreaux L, Werth JL, Poole RM. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain. 2001; 94: 149–158.
Farrar JT, Pritchett YL, Robinson M, Prakash A, Chappell A. The Clinical Importance of Changes in the 0 to 10 Numeric Rating Scale for Worst, Least, and Average Pain Intensity: Analyses of Data from Clinical Trials of Duloxetine in Pain Disorders. J Pain. 2010; 11: 109– 118.
Furquim BD, Flamengui LMSP, Conti PCR. TMD and chronic pain: A current view. Dental Press J Orthod. 2015;20:127–133.
14 Hoffmann RG, Kotchen JM, Kotchen TA, Cowley T, Dasgupta M, Cowley AW.
Temporomandibular disorders and associated clinical comorbidities. Clin J Pain. 2011; 27: 268–274.
Huttunen J, Qvintus V, Suominen AL, Sipilä K. Role of psychosocial factors on treatment outcome of temporomandibular disorders. Acta Odontol Scand. 2019; 77: 119–125.
List T, Jensen RH. Temporomandibular disorders: Old ideas and new concepts. Vol. 37, Cephalalgia. 2017; 37: 692–704.
Nguyen T, Vanichanon P, Bhalang K, Vongthongsri S. Pain Duration and Intensity Are Related to Coexisting Pain and Comorbidities Present in Temporomandibular Disorder Pain Patients. J Oral Facial Pain Headache. 2019; 33: 205–212.
Perrot S, Lantéri-Minet M. Patients’ Global Impression of Change in the management of peripheral neuropathic pain: Clinical relevance and correlations in daily practice. Eur J Pain. 2019; 23: 1117–28.
Salaffi F, Stancati A, Silvestri CA, Ciapetti A, Grassi W. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain. 2004; 8: 283–291.
Salovey P, Smith AF, Turk DC, Jobe JB, Willis GB. The accuracy of memory for pain. Not so bad most of the time. APS J. 1993; 2: 184–191.
Scott W, McCracken LM. Patients’ impression of change following treatment for chronic pain: Global, specific, a single dimension, or many? J Pain. 2015;16:518–526.
Scrivani SJ, Keith DA, Kaban LB. Temporomandibular Disorders. N Engl J Med 2008; 359: 2693-2705.
Shinal RM, Fillingim RB. Overview of Orofacial Pain: Epidemiology and Gender Differences in Orofacial Pain. Dent Clin North Am. 2007; 51: 1–18.
15
Table 1. Diagnoses included in the analyses grouped as Muscular; Arthrogenic (TMJ) and
Other, together with mean symptom severity index score (SSI) at baseline and follow up and percentage change for the different diagnosis groups.
Diagnoses n SSI baseline (mean) SSI follow up (mean) Percent change (mean)
Muscular 25 SD 12 12 SD 12 52% SD 56 Tension-type headache 14 Myalgia 4 Local myalgia 24 Regional myalgia 20 Widespread myalgia 10 TMJ 27 SD 12 8 SD 10 74% SD 31 TMJ dysfunction 5
Discdisplacement with reduction 25
Discdisplacement without reduction 11
Arthralgia 36
Arthritis ass. with underlaying disease or trauma 14
Arthtrosis (local or general) 8
Subluxation 5 Other 25 SD 15 15 SD 14 52% SD 37 Glossopharyngeal neuralgia 1 Tinnitus 1 Pathological attrition 2 Malocclusion 4 Allodynia 1
Headache without specification 5
16
Table 2. The distribution of the percentage change (min, max, mean and SD) in SSI
corresponding to each PGIC-category for the total data and for the muscular and articular diagnoses separately. Only one individual rated the PGIC to “much worse”.
Min SSI% Max SSI% Mean SSI% SD
Total data (all diagnoses)
PGIC -2 = "much worse" 20 %
PGIC 0 = "unchanged" -67 % 94 % 21 % 40 %
PGIC 1 = "minimally improved" -45 % 100 % 45 % 37 % PGIC 2 = "much improved" -275 % 100 % 75 % 46 % PGIC 3 = "vey much imprved" 0 % 100 % 89 % 21 %
Artucular diagnoses
PGIC 0 = "unchanged" 0 % 90 % 38 % 35 %
PGIC 1 = "minimally improved" -20 % 96 % 46 % 34 % PGIC 2 = "much improved" 13 % 100 % 78 % 25 % PGIC 3 = "very much improved" 0 % 100 % 88 % 23 %
Muscular diagnoses
PGIC 0 = "unchanged" -67 % 53 % -6 % 36 %
17
18
Figure 2. Relationship between the mean and 95% confidence interval of reduction of
19
Figure 3. Relationship between the mean and 95% confidence interval of reduction of
20
Appendix
Figure A1. Symptom severity index consisting of frequency of symptoms (0=never; 1=
occasionally; 2= 0nce or twice a month; 3=once a week; 4= several times a week; 5=daily) multiplied with intensity of symptoms (0= no symptoms, 10= unbearable/ maximal symptoms).
Figure A2. Patients global impression of change (-3=very much worse, -2=much worse, -1=
21
Table A1. Multivariate regression results of different variables entered simultaneously in the
equation as potential predictors of impression of change.
