Access to this full-text is provided by Springer Nature.
Content available from BMC Musculoskeletal Disorders
This content is subject to copyright. Terms and conditions apply.
RESEARCH Open Access
This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2025. Open Access This
article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and
reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative
Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence
and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this licence, visit h t t p : / / c r e a t i v e c o m m o n s . o r g / l i c e n s e s / b y / 4 . 0 /.
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
https://doi.org/10.1186/s12891-025-08378-5 BMC Musculoskeletal
Disorders
*Correspondence:
Diane M. Flynn
diane.m.ynn4.civ@health.mil
Full list of author information is available at the end of the article
Abstract
Background The National Institutes of Health (NIH) Research Task Force (RTF) on Research Standards for Chronic
Low Back Pain impact score is a composite measure of Patient Reported Outcomes Measurement Information
System (PROMIS) pain intensity, pain interference and physical function. PROMIS surveys are available in short-form
and computer adaptive testing (CAT) formats. Minimal important change (MIC) can be estimated to determine if
between-group dierences are large enough to be important. To date, three anchor-based estimates of impact
score MIC ranging from 3 to 7.5 have been published, and all were based on data collected using PROMIS short-form
surveys. None used CAT versions of PROMIS surveys.
Methods Secondary analysis of data collected during the conduct of two randomized clinical trials of 6-week
courses of nonpharmacological pain therapies. Research subjects were US active-duty service members referred to
an interdisciplinary pain management center. Impact score was assessed at the beginning and end of treatment.
The Patient Global Impression of Change (PGIC) questionnaire was administered at the end of treatment and asked
respondents to report their status compared to the start of treatment using a 7-item categorical scale ranging from
very much improved to very much worse. A PGIC response of “much” or “very much” improved dened important
improvement. Receiver operating characteristic (ROC) curve analysis and predictive logistic regression models
were used to estimate MIC for the full combined sample and stratied by study sample and baseline impact score.
Measures of individual statistical change were also computed.
Estimating minimal important change
of the National Institutes of health research
task force impact score using computer
adaptive measures: a secondary analysis
of two randomized clinical trials in a military
population with chronic pain
Diane M.Flynn1*, Larisa A.Burke2, Alana D.Steen2, Jerey C.Ransom1, Kira P.Orr3, Honor M.McQuinn1, Tyler
J.Snow1 and Ardith Z.Doorenbos2,4
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 2 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
Background
e pain impact score (PIS)—variably referred to as the
Impact Stratication Score [1–3], RTF impact score [4],
Pain Impact Stratication Score [5], and Pain Impact
Score [6]—is a composite measure of Patient-Reported
Outcomes Measurement Information System (PROMIS)
measures of pain intensity, pain interference, and physi-
cal function. e National Institutes of Health (NIH)
Task Force on Research Standards for Chronic Low Back
Pain (RTF) has endorsed the PIS as a tool to stratify the
impact of musculoskeletal pain on the lives of those who
experience it [1].
e Initiative on Methods, Measurement, and Pain
Assessment in Clinical Trials (IMMPACT) consensus
statement of 2020 states that “statistically signicant evi-
dence of a treatment’s ecacy in a clinical trial is insuf-
cient to indicate that the magnitude of the treatment
eect is clinically important” and that “evaluations of
clinical importance must distinguish between deter-
mining whether the mean improvements are important
to patients or whether the group dierences between
treatments in an RCT are clinically important” [7]. e
statements distinguish between within-patient minimal
clinically meaningful change and between group mini-
mal clinically meaningful dierences. Group level mini-
mal clinically important change (MIC) estimates include
anchor-based approaches that relate change scores on an
instrument to an external criterion of important change.
Distribution based approaches can be used to determine
thresholds for individual statistically signicant change
[8, 9].
It has been argued that anchor-based MIC thresholds
are not appropriate to identify responders to treatment
because MIC estimates are averages derived from distri-
butions of individual MICs and therefore may not reect
the perceived change by a given individual. For that
reason, it is suggested that use of statistical measures of
individual change also be conducted when attempting to
identify responders to treatment [9].
To date, three anchor-based estimates of the PIS MIC
have been published [2–4, 10]. e rst published esti-
mate was based on data collected from a rural population
of adults with musculoskeletal pain, age 55 years or older,
who completed the paper version of the PROMIS-29
questionnaire [4]. e authors recommended further
studies with data collected from other populations, and
advocated for use of computer-adaptive testing (CAT)
versions of the PROMIS measures. CAT is a process by
which sequential items in a questionnaire are based on
previous responses in order to assess the outcome vari-
able using the fewest possible questions. Compared with
the PROMIS 4-item short-form questionnaires, CAT
yields improved accuracy with a slightly higher survey
burden averaging 4.7 items [11]. Yet all three previously
published studies of the PIS MIC collected PIS data using
paper or digital versions of the PROMIS short-form pain
interference and physical function questionnaires; none
collected data using CAT versions of these assessments.
Methods
e primary aim of the study was to address this gap by
estimating the PIS MIC from data collected using CAT
versions of PROMIS measures of pain interference and
physical function. e secondary aim was to explore the
psychometric properties of the PIS. e data used for this
secondary analysis were collected during two unrelated
randomized clinical trials (RCTs) of nonpharmacological
pain therapies provided to US active-duty Army, Navy,
and Air Force service members. Both clinical trials were
approved by the Madigan Army Medical Center insti-
tutional review board. e Integrative Modalities Plus
Psychological, Physical, and Occupational Restorative
Results Overall, a decrease of 3 points in impact score was the estimated MIC (2.5 for ROC analysis and 3.4 for
predictive modeling approach). Larger decreases in impact score were needed for participants with moderate and
severe baseline pain impact to report important improvement. Thresholds for individual statistically signicant change
ranged from 6 to 14.
Conclusions Using data collected with CAT surveys, we calculated an MIC of 3 points for the NIH RTF impact score,
and estimates ranged from 1.3 to 7.2 depending on the baseline impact score and statistical approach used. These
ndings are consistent with previous MIC estimates that were based on non-adaptive short form surveys and have
implications for improving the accuracy of pain treatment response assessment.
Registry information Trial registration. ClinicalTrials.gov. Registry numbers: NCT03297905 (registered 9/29/17) and
NCT04656340 (registered 11/30/20). Link to full applications: h t t p s : / / c l a s s i c . c l i n i c a l t r i a l s . g o v / c t 2 / s h o w / N C T 0 3 2 9 7 9 0 5
? t i t l e s = D e t e r m i n a n t s + o f + O p t i m a l + D o s a g e % 2 6 c n t r y = U S % 2 6 d r a w = 2 % 2 6 r a n k = 1; h t t p s : / / c l a s s i c . c l i n i c a l t r i a l s . g o v / c t 2 /
s h o w / r e s u l t s / N C T 0 4 6 5 6 3 4 0 ? t i t l e s = C o m p l e m e n t a r y + a n d + I n t e g r a t i v e + p a i n + t h e r a p i e s + a n d + f u n c t i o n a l + r e s t o r a t i o n +
% 2 8 I M P P P O R T % 2 9 % 2 6 d r a w = 2 % 2 6 r a n k = 1. Patient enrollment dates: SMART: 17 March 2021, prospectively registered;
IMPPPORT: 9 December 2015, retrospectively registered.
Keywords Chronic pain impact, PROMIS, Minimal important change
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 3 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
erapies (IMPPPORT) study enrolled 210 participants
and was designed to determine if adding complemen-
tary and integrative health therapies enhanced the out-
comes of an intensive functional restoration program
(protocol #215050). e Complementary, Integrative,
and Standard Rehabilitative Pain erapies Pragmatic
Trial with SMART Design (SMART study) enrolled 280
participants and was designed to determine the optimal
sequence, duration, and combination of physical, occu-
pational, and complementary and integrative health
therapies for treatment of chronic, predominantly mus-
culoskeletal pain (protocol #221011). e research proto-
cols for both RCTs have been described in detail [12, 13].
Study participants
Research participants for both RCTs were recruited
from the population of US active-duty service members
referred to the Madigan Army Medical Center Interdis-
ciplinary Pain Management Center. Enrollment occurred
between 9 December 2015 and 16 May 2018 for the
IMPPPORT study and between 17 March 2021 and 9
September 2022 for the SMART study. Both studies’
inclusion criteria required active-duty military status and
functional impairment due to pain. In addition, inclusion
criteria for the IMPPPORT study required the ability to
meet modest functional thresholds (stand up from and
sit down on oor independently, walk or jog on a tread-
mill for at least 6min, lift and/or carry at least 20 lbs.)
to ensure participants’ capacity to engage in an intensive
functional restoration program. Exclusion criteria for
both studies were inability to commit the time required
for treatment, surgery within the previous or following
6 months, unstable psychological condition(s) or being
in the process of medical disability determination at the
time of study enrollment. e combined enrolled popula-
tion of both RCTs was 365; no participants were enrolled
in both studies.
Study procedures
For each of the original RCTs, potential participants were
recruited through in-person discussion with a member of
the research team immediately following a routine inter-
disciplinary pain management center visit. No compen-
sation was oered for participation. Following informed
consent, participants in each study were randomized to
the rst of two 3-week treatment stages. In both studies,
participants were asked to provide measures at baseline,
at the end of stage 1 and stage 2, and at 3- and 6-month
post-treatment follow-ups. In addition, participants were
asked at the end of stage 2 to report their perceived level
of improvement relative to baseline.
Study measures
Pain impact score (PIS) components
e PIS was originally described as the sum of the PRO-
MIS 7-day average pain intensity (from 0 = “no pain”
through 10 = “worst imaginable pain”), PROMIS pain
interference short-form 4a v1.1 score (range: 4–20),
and the reverse of the PROMIS physical function short-
form 4a v.1.0 score (range: 4–20). is computation
resulted in a PIS that ranged from 8 (least impact) to
50 (most impact) [4]. Our current analysis instead used
the Defense and Veterans Pain Rating Scale (DVPRS) to
determine 7-day average pain intensity and CAT versions
of the PROMIS pain interference [14, 15] and physical
function measures [16]. e DVPRS also has a range of
0 (“no pain”) to 10 (“as bad as it could be, nothing else
matters”); uses a combination of numeric, color, facial
expression, and word descriptors; and has been validated
in military and veteran populations [17, 18]. e CAT
versions of PROMIS report T-scores; thus, to compute
the PIS, we took the additional step of converting the
PROMIS T-scores into scores equivalent to the PROMIS
short-form scores. T-score equivalents of each short-
form score are published in the user manuals for PRO-
MIS pain interference and physical function measures
[19]. Steps for calculating pain impact score using PRO-
MIS computer adaptive tests are described in Table1.
Additional PROMIS measures
In addition to the PIS components described above, this
study’s outcome measures included the CAT versions of
PROMIS sleep-related impairment, fatigue, depression,
anxiety, and anger questionnaires [20–22].
Patient global impression of change
e Patient Global Impression of Change (PGIC) ques-
tionnaire [23] asks respondents to report their overall
status compared to their status at the start of the research
study, using the following 7-item categorical scale:
“very much improved,” “much improved,” “minimally
improved,” “no change,” “minimally worse,” “much worse,”
“very much worse.” All participants in the SMART study
were asked to complete the PGIC at the end of stage 2
of their study treatment and, in addition, to use the same
7-item scale to report perceived change in each of the 3
components of the PIS: (a) 7-day average pain intensity,
(b) physical function (phrased as “ability to engage in
physical activities”), and (c) pain interference (phrased as
“ability to do day-to-day tasks and enjoy recreational and
social activities”). e IMPPPORT study added collec-
tion of posttreatment PGIC questionnaires in June 2017
through an IRB-approved protocol modication, then
collected PGIC data from the nal 85 of 210 participants
to enroll.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 4 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
Statistical methods
One recommended approach to anchor-based MIC esti-
mation is using receiver operating characteristic (ROC)
curves to determine the optimal cutpoint between
change in an outcome measure and the corresponding
self-report of perceived improvement, such as “much
improved.” [8] ere is no consensus on which PGIC
response corresponds with both “minimal” and “impor-
tant” change, but previous PIS MIC analyses judged
responses of “little” improvement to be insucient,
and used cutpoints of at least “moderately” or “much”
improved to indicate a MIC [2, 4, 10]. Unlike previous
studies of PIS MIC estimates, the version of PGIC used
in the current analysis did not include the category of a
“little” improvement; the closest match is “minimally
improved”. In an eort to be as consistent as possible with
previous research, we considered “minimally improved”
to be insucient to be important to most respondents
and used “much improved” as the cutpoint for MIC.
A disadvantage of the ROC approach is that the MIC
estimate it yields will be an under- or overestimate if
the percentage who report important improvement is
not equal to 50%.8 When this percentage is consider-
ably larger or smaller than 50%, a preferred anchor-based
method of estimating MIC is predictive modeling using
logistic regression. With this method, the dichotomous
variable indicating improvement is the outcome, the
change score is the predictor, and the MIC value is the
change score associated with a likelihood ratio of 1 [24].
Baseline characteristics were assessed separately for
each study population (IMPPPORT and SMART) and
compared statistically with t-test, chi-square, and Fisher’s
exact tests where applicable. Because the PGIC items
were used as anchors for MIC, we compared the relation-
ship between the mean change in PIS (pre to post) and
the PGIC items for each study population by calculating
bivariate correlations stratied by study and graphing
the relationships with box plots. PGIC response catego-
ries were collapsed into bivariate categories for the MIC
analysis, which compared either “much improved” or
“very much improved” to other categories representing
less or no improvement or worsening status. e pri-
mary anchor for the MIC analysis was change in overall
status. We also conducted MIC analysis for the SMART
study using the PGIC items on pain intensity, physical
function, and pain interference as anchors. For the ROC
approach, the CUTPT module in Stata (StataCorp LLC,
College Station, Texas, USA) [25] was used to estimate
anchor-based MIC, and the Euclidean distance method
(point nearest to 0,1) was used to estimate the cutpoint
that maximized both sensitivity and specicity [26]. For
the predictive modeling approach, the MIC values were
calculated using a spreadsheet included as a supplement
to Terluin et al. (2015) [24]. Values were adjusted for the
proportion of patients reporting important improve-
ment. Statistical measures used to assess signicance of
individual change included Standard Deviation Index,
Standard Error of Measurement, Standard Error of Esti-
mate, Standard Error of Prediction, Reliable Change
Index and the Coecient of Repeatability [9].
Responsiveness is the ability of a measure to detect
clinical changes [2]. e responsiveness of the PIS was
assessed by the presence of signicant (at P <.05) mean
change pre to post intervention as determined by paired
t-tests and by the area under the curve (AUC) values
from the ROC curve analysis. An AUC higher than 0.70
was considered responsive as it indicates a high correla-
tion between change in PIS and patient reported change.
We assessed MICs and responsiveness of the PIS for the
combined sample, for each study separately, and strati-
ed by baseline impact score.
Table 1 Calculation of pain impact score from PROMIS
computer adaptive instruments
Physical function 4a - Adult v1.0 Pain interference 4a - Adult
v1.1
Raw score T-score Raw
score
T-score
Low cut-o High
cut-o
Low cut-o High
cut-o
4≥ 0 < 22.9 4≥ 0 < 41.6
4≥ 22.9 < 26.9 4≥ 41.6 < 49.6
5≥ 26.9 < 29.1 5≥ 49.6 < 52.0
6≥ 29.1 < 30.7 6≥ 52.0 < 53.9
7≥ 30.7 < 32.1 7≥ 53.9 < 55.6
8≥ 32.1 < 33.3 8≥ 55.6 < 57.1
9≥ 33.3 < 34.4 9≥ 57.1 < 58.5
10 ≥ 34.4 < 35.6 10 ≥ 58.5 < 59.9
11 ≥ 35.6 < 36.7 11 ≥ 59.9 < 61.2
12 ≥ 36.7 < 37.9 12 ≥ 61.2 < 62.5
13 ≥ 37.9 < 39.1 13 ≥ 62.5 < 63.8
14 ≥ 39.1 < 40.4 14 ≥ 63.8 < 65.2
15 ≥ 40.4 < 41.8 15 ≥ 65.2 < 66.6
16 ≥ 41.8 < 43.4 16 ≥ 66.6 < 68.0
17 ≥ 43.4 < 45.3 17 ≥ 68.0 < 69.7
18 ≥ 45.3 < 48.0 18 ≥ 69.7 < 71.6
19 ≥ 48.0 < 56.9 19 ≥ 71.6 < 75.6
20 ≥ 56.9 <=100 20 ≥ 75.6 <= 100
Step 1. Convert pain interference and physical function CAT T-Scores to short-
form raw score s using conversion table s above
Step 2: Reverse p hysical function raw s core by subtracting it fr om 24
Step 3: Add raw score for pain interference and reversed raw score for physical
functio n to the 7-day average pain inte nsity score (patien t rating from 0 to 10 of
average pain in tensity in the past 7 days)
Example: Average pain intensity is 7. Pain interference T-score is 60.10, raw
score equivalent is 11. Physical function T-score is 35.05, raw score equivalent
is 10, reverse raw scor e is 24 − 10 = 14. PIS = 7 + 11 + 14 = 32
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 5 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
Results
e sample for this secondary analysis included 61 par-
ticipants from the IMPPPORT study and 192 partici-
pants from the SMART study who had completed the
posttreatment PGIC and had both pre- and posttreat-
ment PIS scores. e demographic characteristics, clini-
cal characteristics, and PGIC responses of both study
populations are shown in Table2. e combined popu-
lation (n = 253) was predominantly married male Army
service members over 25 years of age with at least some
college education. Compared to the IMPPPORT study
population, the SMART study population was older, of
higher military rank, were less likely to have their princi-
pal pain condition coded as musculoskeletal type and had
a higher (worse) baseline PIS; higher (worse) PROMIS
scores for baseline pain interference, depression, anger,
and sleep impairment; and a lower (worse) PROMIS
score for physical function. For both studies, the most
common responses for improvement in overall status
were “minimally improved” (36%) and “much improved”
(32%); while 6% reported “very much improved” and 24%
reported either no change or a worsening from baseline.
Figure 1 displays the SMART study participants’
responses to all four PGIC items. Reported impression
of change varied across items. Results for Fisher’s exact
tests found statistical dierences between reported
overall change compared to physical activity (P =.01)
and ability to do daily tasks (P =.03). A greater propor-
tion of participants reported “minimal” or no change in
improvement in physical activity compared to overall
status (71% vs. 52%, P =.01). Also, a smaller proportion of
participants reported “much improvement” in the ability
to do daily tasks (25% vs. 32%) and a greater proportion
of participants reported “no change” (30% vs. 16%) com-
pared to overall status. e association between overall
PGIC status and change in PIS was slightly stronger for
the SMART study sample compared to the IMPPPORT
sample (r = 0.6 vs. r = 0.4, P =.09), but the dierence was
not statistically signicant.
Table3 shows pre-treatment PIS, post-treatment PIS,
mean and percentage of PIS change during treatment,
and mean PIS change by self-reported change in over-
all status for the combined sample as well as stratied
by study and by baseline PIS (mild, moderate or severe).
For the combined sample, the average change in PIS
was a decrease of 2.7 points, and the average percent-
age change was a decrease of 8.5%. ese changes repre-
sented a statistically signicant decrease (at P <.05). e
mean change and percentage change in PIS did not dif-
fer between studies. e mean change in PIS was smaller
for the group with a mild baseline PIS compared to the
groups with moderate or severe baseline PIS; similarly,
the mean percentage change was smaller for the mild
compared to the moderate PIS group.
Results from the ROC and predictive modeling analy-
ses are shown in Table 3. e overall recommended
MIC for the combined sample was a decrease of 3 in PIS
(2.5 for ROC analysis and 3.4 for the predictive model-
ing approach). Figure2 displays the distribution of PIS
change scores above and below the MIC of 3 for partici-
pants who did and did not report important improve-
ment. ROC curve is provided in Supplemental Fig. 1.
e sensitivity and specicity for the ROC analysis were
Table 2 Participant demographic and clinical characteristics at
Baseline and impressions of Change following treatment
IMPPPORT
MIC popula-
tion (n = 61)
SMART MIC
population
(n = 192)
Sig
di
Age (years) n%n% *
<25 17 27.9% 13 6.8%
25–34 21 34.4% 88 45.8%
≥35 23 37.7% 91 47.4%
Malea52 85.2% 150 78.1%
Married/partnered 44 72.1% 136 70.8%
High school/GED onlyb19 31.1% 45 23.4%
Active-duty Army 52 85.2% 158 82.3%
Military rank: junior enlisted 25 41.0% 50 26.0% *
Pain type *
Musculoskeletal (ICD-10
M00-M99)
45 86.3% 122 64.5%
Nerves and senses (ICD-
10 G00-H95, includes G89.xx
chronic pain codes)
6 9.8% 60 31.3%
Other 2 3.9% 10 5.2%
PROMIS mean T-scores dMean SD Mean SD
Pain interference 61.3 4.4 64.5 5.3 *
Physical function 41.6 5.1 39.7 5.0 *
Depression 50.0 9.2 54.7 10.4 *
Anxiety 53.9 10.4 56.7 9.9
Anger 50.4 10.8 56.6 11.8 *
Fatigue 57.4 10.1 59.9 9.3
Sleep-related impairment 58.7 9.6 61.9 9.2 *
Pain intensity 5.5 1.2 5.6 1.5
Pain impact score 26.4 6.0 29.9 7.0 *
Overall status
Very much improved 7 11% 8 4%
Much improved 21 34% 60 32%
Minimally improved 21 34% 69 36%
No change 7 11% 33 17%
Minimally worse 5 8% 12 6%
Much worse 0 0% 7 4%
a The demographic questionnaires oered only two choices for sex: male or
female
b One part icipant was missing data on e ducation in IMPPPORT s tudy
c Pain type w as missing for 10 IMPPPORT subje cts
d PROMIS measur es are reported b y T-scores, with a range of 0 -100, US reference
populatio n mean of 50.0 and SD equal to 10. For phy sical function, lo wer scores
indicate worse functioning relative to higher scores. For all other PROMIS
measures lis ted, higher scores indic ate worse scores relative t o lower scores
* Signican ce was set at P <.05
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 6 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
76% and 66% respectively. With the exception of the MIC
specic to the IMPPPORT study subgroup, all AUCs
exceeded 0.70. e MIC estimates were smaller for those
with mild baseline PIS and larger for those with moderate
or severe baseline PIS. e MIC estimates were smaller
in the IMPPPORT study sample than in the SMART
study sample. For the SMART study, using pain inten-
sity, ability to engage in physical activities, or ability to do
Table 3 Responsiveness and minimal important change in Pain Impact score, by study and baseline impact score
Study Category of baseline PIS
Imppport Smart Mild
(8–27)
Moderate
(28–34)
Severe
(35–50)
Overall
Participants, n61 192 107 89 57 253
Scores, mean (SD)
Pre-treatment PIS 26.4 (6.0) 29.9 (7.0) 22.6 (3.3) 30.7 (1.9) 38.7 (3.4) 29.1 (6.9)
Post-treatment PIS 23.9 (6.1) 27.1 (8.8) 21.4 (6.2) 27.1 (6.1) 34.6 (8.1) 26.4 (8.3)
Mean change in PIS −2.4 (6.2) −2.8 (6.4) −1.2 (5.8) −3.6 (6.0) −4.0 (7.4) −2.7 (6.4)
Percentage change in PIS −7.1 (23.7) −9.0 (22.4) −4.7 (26.0) −11.8 (19.4) −10.5 (19.8) −8.5 (22.7)
Mean change in PIS by self-reported overall status
Very much improved -10 (5.9) -10 (7.7) -7.1 (7.1) -10 (4.6) -19 (1.4) -9.9 (6.7)
Much improved -3.3 (5.4) -6.7 (5.5) -3.6 (4.9) -7.7 (4.6) -10.9 (6.3) -5.8 (5.6)
Minimally improved -0.9 (5.2) -2.5 (5.0) 0.8 (4.1) -3.4 (4.2) -4.8 (5.6) -2.2 (5.1)
No change -0.3 (6.1) 1.5 (5.1) 2.4 (5.8) 1.1 (3.7) -0.3 (6.0) 1.2 (5.3)
Minimally worse 2.6 (6.2) 1.4 (4.2) 2.8 (7.2) 4.4 (3.6) -0.4 (3.2) 1.8 (4.7)
Much worse 5.6 (5.2) 4 8 (7.1) 4.8 (5.7) 5.6 (5.3)
MIC ROC approach
Participants, n61 189 106 88 56 250
Cutpoint -2.5 -4.5 -0.5 -4.5 -8.5 -2.5
95% CI -5.5, 0.5 -7.1, -1.9 -2.1, 1.1 -6.4, -2.6 -14.5, -2.4 -5.0, 0.0
AUC 0.67 0.74 0.75 0.78 0.77 0.71
Sensitivity 68% 69% 81% 77% 75% 76%
Specicity 67% 79% 70% 79% 80% 66%
MIC predictive modeling approach
Cutpoint -3.5 -3.8 -1.3 -4.7 -7.2 -3.4
95% CI -15.2, -0.1 -9.4, -1.2 -5.3, 0.7 -8.5, -2.5 -12.3, -4.4 -10.1, -0.7
% much or very much improved 49% 36% 50% 35% 21% 38%
Legend: AUC = area under the curve; CI, condence interval; MIC = minimal important change; PIS, pain impact score; ROC = receiver operating characteristic.
Cutpoint fro m Euclidean distance met hod (nearest to 0,1) displayed. Anchor for R OC analysis is PGIC item fo r patient-reporte d improvement in overall sta tus
Fig. 1 Patient impression of change following treatment. Distribution of 7 oered responses ranging from “very much improved” to “very much worse”
for each anchor: pain intensity, physical function, pain interference and overall status
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 7 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
day-to-day tasks as anchors—as compared to improve-
ment in overall status—did not result in notably dierent
MIC estimates.
Statistical measures used to assess signicance of
individual change were as follows: Standard Deviation
Index = 14, Standard Error of Measurement = 7, Standard
Error of Estimation = 6, Standard Error of Prediction = 9,
and Coecient of Repeatability (individual) = 9.1. e
Coecient of Repeatability for group level change was
0.6 [9].
Discussion
e key nding from this analysis was that when using
CAT versions of PROMIS physical function and pain
interference, the PIS MIC corresponding to “much” or
“very much” improvement was a decrease of about 3
points, depending on statistical approach (2.5 for ROC
vs. 3.4 for predictive modeling). Given our observation
of 38% of participants who reported more than mini-
mal change and given the bias introduced in the ROC
approach when the percentage is not equal to 50%, 3.4
is likely the more precise estimate. Statistics for signi-
cance of individual change were notable higher, ranging
from 6 to 14 indicating that a larger change in PIS score
is needed to conrm that an individual patient’s change is
not due to measurement error. e MIC estimate varied
substantially by baseline impact score, demonstrating
that patients with greater initial pain-related limitations
may need to experience greater improvements in pain
intensity, pain interference and/or physical function
before meeting the threshold for important change in
overall status. is dependence of MIC on baseline values
may support the use of dierent methods for categoriz-
ing important improvement, such as meeting a specied
threshold (e.g., decreasing from moderate or severe pain
impact to mild impact). e MIC also diered between
the two study samples, likely due at least in part to a
higher mean impact score in the SMART study sample.
A continuing challenge in estimating the PIS MIC is
the lack of consensus on preferred methodology, such
as (a) which anchor is assessed (e.g., pain, overall sta-
tus); (b) number and categories of improvement options
oered; and (c) threshold at which change is perceived
as “important” (e.g., “minimal” vs. “much” improvement)
[9]. Table4 summarizes these dierences in previously
published MIC estimates. Overall, our MIC estimates are
consistent with previous estimates, which range from 3
to 7.5: Our estimate mirrors that of Deyo et al., who esti-
mated the MIC at 3, using a 5-option scale in an older
population of rural primary care patients [4], but is some-
what lower than that of Dutmer et al., who estimated the
MIC at 7.5 in a population with chronic low back pain
Fig. 2 Distribution of change in pain impact scores above and below the minimal important change of 3 for patients that did and did not report im-
provement in overall status
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 8 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
cared for in a Dutch spine center [10]. In similar ndings,
Hays et al. estimated a MIC of 7 in a military population
seeking chiropractic care for low back pain of any dura-
tion, using a 7-item categorical scale; unlike other stud-
ies, the scale used by Hays et al. included the option of
“moderately improved.” [2].
Our study has important implications particularly for
researchers who want to examine PIS MIC using CAT
versions of PROMIS measures. Researchers proposing
to use PIS as an outcome measure can consider between-
group decrease of at least 3–7 points to indicate impor-
tant group-level dierence, but should be aware that
individual change scores of 14 or fewer points may be
due to measurement error and should therefore be inter-
preted with caution. Additional research is needed to
determine if using a specic PIS cutpoint, such as scores
in the “mild” range (i.e., equal to or less than 27 points
[1]), rather than an absolute change in PIS, is a better
approach to dening response. In clinical practice, our
ndings can guide clinicians in determining response
to treatment and making decisions on whether current
treatment should be continued or modied.
A limitation of this study is that the 95% condence
limits for the overall MIC estimates yielded by the ROC
and predictive modeling approaches included ranges
from no change to 10.1 points improvement, thus future
use of MIC estimates should also consider population
characteristics and baseline PIS. In addition, our sample
included only active-duty service members, and thus may
not be representative of nonmilitary populations. Also,
we used the DVPRS rather than the PROMIS pain inten-
sity scale to compute PIS, which may have had modest
eects on the impact score.
Other limitations include the lack of consensus about
whether anchor-based MICs should be used to identify
individual responders to treatment. Terwee and col-
leagues [8] suggest that MIC estimates can be used to
identify responders to treatment. Conversely, Hays and
Peipert [9] argue that MIC thresholds should not be used
to identify responders to treatment because MIC esti-
mates are averages derived from distributions of indi-
vidual MICs and therefore may not reect the perceived
change by a given individual. Hays and Peipert recom-
mend presenting both individual statistical signicance
and whether the individual feels they have improved to
identify responders to treatment. In addition, Hays and
Peipert note that including all those who change rather
than focusing on those with minimal but important
change can result in the MIC threshold estimates that are
too large.
Conclusions
We used the CAT versions of PROMIS physical function
and pain interference measures to obtain an estimate of
the PIS MIC that was similar to those reported in pre-
vious studies that used PROMIS short-form measures.
Table 4 Prior research that estimated minimal important change in pain impact score
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 9 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
is similarity opens new options for researchers and
clinicians when assessing pain treatment response to use
CAT questionnaires, which have greater accuracy than
short-form questionnaires.
Abbreviations
AUC Area under the curve
CAT Computer adaptive testing
CI Condence interval
CUTPT Module to determine optimal cutpoint in Stata statistical
software package
DVPRS Defense and Veterans Pain Rating Scale
GED General education diploma
ICD-10 International classication of diseases, 10th edition
IMPPPORT Integrative Modalities Plus Psychological, Physical and
Occupational Restorative Therapies
IRB Institutional review board
LBP Low back pain
MIC Minimal important change
NIH RTF National Institutes of Health Task Force on Research Standards
for Chronic Low Back Pain
PGIC Patient global impression of change
PIS Pain impact score
PROMIS Patient-Reported Outcomes Measurement Information System
RCI Reliable change index
RCT Randomized clinical trial
ROC Receiver operating characteristic
SMART Sequential multiple assignment randomized trial
US United States
Supplementary Information
The online version contains supplementary material available at h t t p s : / / d o i . o r
g / 1 0 . 1 1 8 6 / s 1 2 8 9 1 - 0 2 5 - 0 8 3 7 8 - 5.
Supplementary Material 1: Supplemental Fig.1. Receiver operating char-
acteristic curve for impact score change cutpoint associated with “much
improved” or “very much improved” overall status
Acknowledgements
The authors wish to extend their appreciation to Norma Bowling of Kennell
and Associates, in support of the Defense Health Agency Enterprise
Intelligence and Data Solutions Program Management Oce, for her data
management support. In addition, Alexandra Fairchok and James Martinez
provided administrative support during the conduct of the study, and Kyra
Freestar of Bridge Creek Editing provided professional editing support.
Author contributions
D.F. conceived of the analysis and drafted the research protocol pertaining
to this analysis. Served as protocol principal investigator. L.B. conducted the
statistical analysis. A.S. provided input into statistical analysis. D.F. and L.B.
drafted the manuscript and prepared tables and gures. K.O. coordinated
study conduct and data collection. Served as research coordinator. H.M., T.S.
and J.R. assisted with study conduct and data collection. A.D. served as overall
grant PI. All authors participated in weekly research meetings during the
study period, and provided input into interpretation of results and clinical and
research implications. All authors reviewed and approved the manuscript.
Funding
The US Army Medical Research Acquisition Activity, 820 Chandler Street, Fort
Detrick, MD 21702 − 5014, is the awarding and administering acquisition
oce. This work was supported by the Assistant Secretary of Defense
for Health Aairs endorsed by the Department of Defense through the
Neuromusculoskeletal Injuries Rehabilitation Research Award under Award
No. W81XWH-18-2-0023, and through the Clinical Research Initiative (CRI)
Intramural Research Award under Award No. W81XWH-14-DMRDP-CRI-IRA-
MTI. This study was also supported by funding by the National Institutes of
Health/National Institute of Neurological Disorders and Stroke under Award
No. K24 AT011995.
Data availability
The current approved research protocol does not permit sharing of data that
were used for this analysis. However, reasonable requests for de-identied
data to the corresponding author may be considered for release following an
IRB-approved protocol modication.
Declarations
Ethics approval and consent to participate
The study protocol and informed consent form were determined to meet the
revised Common Rule by the Madigan Army Medical Center Human Research
Protections Oce and was approved by the Madigan Army Medical Center
Institutional Review Board, protocol number 218052. No nonhuman animal
species were used in the conduct of this research.
Consent for publication
Not applicable.
Disclaimer
The views expressed are those of the authors and do not reect the policy
or position of the Department of the Army, Department of Defense, the
National Institutes of Health or the US Government. The content is solely the
responsibility of the authors. The investigators adhered to the policies for
protection of human subjects as prescribed in 45 CFR 46.
Competing interests
The authors declare no competing interests.
Author details
1Interdisciplinary Pain Management Center, Madigan Army Medical
Center, 9040 Jackson Avenue, Tacoma, WA 98431, USA
2College of Nursing, University of Illinois Chicago, 845 S. Damen Avenue,
Chicago, IL 60612, USA
3The Geneva Foundation, 950 Broadway, Suite 307, Tacoma, WA
98402, USA
4Department of Anesthesiology and Pain Medicine, University of
Washington School of Medicine, 1959 NE Pacic St., Campus Box 356540,
Seattle, WA 98195, USA
Received: 27 December 2023 / Accepted: 30 January 2025
References
1. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on
Research Standards for Chronic Low Back Pain. Pain Med. 2014;15(8):1249–67.
h t t p s : / / d o i . o r g / 1 0 . 1 1 1 1 / p m e . 1 2 5 3 8.
2. Hays RD, Orlando Edelen M, Rodriguez A, Herman P. Support for the Reliabil-
ity and Validity of the National Institutes of Health Impact Stratication Score
in a sample of active-duty U.S. Military Personnel with Low Back Pain. Pain
Med Malden Mass. 2021;22(10):2185–90. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / p m / p n a b 1 7
5.
3. Hays RD, Slaughter ME, Spritzer KL, Herman PM. Assessing the Signicance
of Individual Change in 2 samples of patients in treatment for low back
Pain using 5 dierent statistical indicators. J Manipulative Physiol Ther.
2021;44(9):699–706. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . j m p t . 2 0 2 2 . 0 3 . 0 0 2.
4. Deyo RA, Katrina Ramsey null, Buckley DI et al. Performance of a Patient
Reported Outcomes Measurement Information System (PROMIS) Short Form
in Older Adults with Chronic Musculoskeletal Pain. Pain Med Malden Mass.
2016;17(2):314–324. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / p m / p n v 0 4 6
5. Angarita-Fonseca A, Pagé MG, Meloto CB, et al. The Canadian version of
the National Institutes of Health minimum dataset for chronic low back
pain research: reference values from the Quebec Low Back Pain Study. Pain.
2023;164(2):325–35. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 7 / j . p a i n . 0 0 0 0 0 0 0 0 0 0 0 0 2 7 0 3.
6. Pope JE, Fishman M, Chakravarthy K, et al. A retrospective, Multicenter,
quantitative analysis of patients’ Baseline Pain Quality (PROMIS-29) enter-
ing into Pain and Spine practices in the United States (ALIGN). Pain Ther.
2021;10(1):539–50. h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 4 0 1 2 2 - 0 2 1 - 0 0 2 3 8 - z.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Page 10 of 10
Flynn et al. BMC Musculoskeletal Disorders (2025) 26:137
7. Smith SM, Dworkin RH, Turk DC, et al. Interpretation of chronic pain
clinical trial outcomes: IMMPACT recommended considerations. Pain.
2020;161(11):2446–61. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 7 / j . p a i n . 0 0 0 0 0 0 0 0 0 0 0 0 1 9 5 2.
8. Terwee CB, Peipert JD, Chapman R, Lai JS, Terluin B, Cella D, Grith P, Mok-
kink LB. Minimal important change (MIC): a conceptual clarication and
systematic review of MIC estimates of PROMIS measures. Qual Life Res.
2021;30:2729–54.
9. Hays RD, Peipert JD. Between-group minimally important change versus
individual treatment responders. Qual Life Res. 2021;30(10):2765–72.
10. Dutmer AL, Reneman MF, Schiphorst Preuper HR, Wol AP, Speijer BL, Soer
R. The NIH minimal dataset for chronic low back Pain: responsiveness and
minimal clinically important change. Spine. 2019;44(20):E1211–8. h t t p s : / / d o i .
o r g / 1 0 . 1 0 9 7 / B R S . 0 0 0 0 0 0 0 0 0 0 0 0 3 1 0 7.
11. Segawa E, Schalet B, Cella D. A comparison of computer adaptive tests (CATs)
and short forms in terms of accuracy and number of items administrated
using PROMIS prole. Qual Life Res. 2020;29:213–21.
12. Flynn DM, McQuinn H, Burke L, Steen AD, Fairchok A, Snow T, Doorenbos
AZ. Use of complementary and integrative health therapies before intensive
functional restoration in active-duty service members with chronic pain. Pain
Med. 2022;23:844–56. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / p m / p n a b 3 2 6.
13. Flynn DM, Eaton LH, Langford DJ, Ieronimakis N, McQuinn H, Burney RO,
Holmes SL. Doorenbos AZ. A SMART design to determine the optimal
treatment of chronic pain among military personnel. Contemp Clin Trials.
2018;73:68–74.
14. Amtmann D, Cook KF, Jensen MP, et al. Development of a PROMIS item bank
to measure pain interference. Pain. 2010;150(1):173–82.
15. Askew RL, Cook KF, Revicki DA, Cella D, Amtmann D. (2016). Clinical Validity of
PROMIS® pain interference and pain behavior in diverse clinical populations. J
Clin Epidemiol. 2016 May;73:103-111.
16. Schalet BD, Hays RD, Jensen SE, Beaumont JL, Fries JF, Cella D. Validity of PRO-
MIS® physical function measures in diverse clinical samples. J Clin Epidemiol.
2016;73:112–8.
17. Polomano RC, Galloway KT, Kent ML, et al. Psychometric testing of the
Defense and Veterans Pain Rating Scale (DVPRS): a new pain scale for military
population. Pain Med Malden Mass. 2016;17(8):1505–19. h t t p s : / / d o i . o r g / 1 0 . 1 0
9 3 / p m / p n w 1 0 5.
18. Buckenmaier CC 3rd, Galloway KT, Polomano RC, McDue M, Kwon N, Galla-
gher RM. Preliminary validation of the Defense and Veterans Pain Rating Scale
(DVPRS) in a military population. Pain Med. 2013;14(1):110–23. h t t p s : / / d o i . o r g
/ 1 0 . 1 1 1 1 / j . 1 5 2 6 - 4 6 3 7 . 2 0 1 2 . 0 1 5 1 6 . x. Epub 2012 Nov 8. PMID: 23137169.
19. HealthMeasures. PROMIS user manuals-pain interference, physical function.
HealthMeasures. Accessed December 15. 2023. h t t p s : / / w w w . h e a l t h m e a s u r e s .
n e t / s c o r e - a n d - i n t e r p r e t / i n t e r p r e t - s c o r e s / p r o m i s
20. Cella D, Riley W, Stone A, et al. The patient-reported outcomes Measure-
ment Information System (PROMIS) developed and tested its rst wave of
adult self-reported health outcome item banks: 2005–2008. J Clin Epidemiol.
2010;63(11):1179–94.
21. Pilkonis PA, Choi SW, Reise SP, et al. Item banks for measuring emotional dis-
tress from the patient-reported outcomes Measurement Information System
(PROMIS®): depression, anxiety, and anger. Assessment. 2011;18(3):263–83.
22. Yu L, Buysse DJ, Germain A, et al. Development of short forms from the
PROMIS™ sleep disturbance and sleep-related impairment item banks. Behav
Sleep Med. 2012;10(1):6–24.
23. Guy W, editor. ECDEU assessment manual for psychopharmacology. U.S.
Department of Health, Education, and Welfare; 1976.
24. Terluin B, Eekhout I, Terwee CB, de Vet HC. Minimal important change (MIC)
based on a predictive modeling approach was more precise than MIC based
on ROC analysis. J Clin Epidemiol. 2015;68(12):1388–96.
25. Clayton P. CUTPT: Stata module for Empirical estimation of cutpoint for a
diagnostic test. Published online 2013. Accessed December 15, 2023. h t t p s : / /
E c o n P a p e r s . r e p e c . o r g / R e P E c : b o c : b o c o d e : s 4 5 7 7 1 9
26. Nahm FS. Receiver operating characteristic curve: overview and practical use
for clinicians. Korean J Anesthesiology. 2022;75(1):25–36.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional aliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
