General health (PROMIS)

The Patient Reported Outcomes Measurement Information System-29 Profile v1.0 (PROMIS-29) (section: diseases & symptoms) is a collection of seven short forms measuring the following seven domains of self-reported health in 4 items each:

1. depression
2. anxiety
3. physical function
4. pain interference (including a 5th item on pain intensity)
5. fatigue
6. sleep disturbance
7. satisfaction with participation in social roles

The PROMIS-29 for adults was used in Lifelines general assessment 2B and in additional assessment ACTQ.
Underage participants received the 10/8-item PROMIS short forms on fatigue and social roles.

The PROMIS-29 Profile is part of a large system of item banks and instruments that have been validated extensively in the US. Each item bank was developed and validated separately. Subsequently, short forms were developed for each bank, which were combined in the PROMIS-29 Profile.
The short forms that make up the PROMIS Profile Instruments include “high information” items. The selection of items was in part based on item rankings within each domain using two psychometric criteria: (1) maximum interval information, and 2) Computer Adaptive Testing (CAT) simulations.
The PROMIS-29 Profile has been translated into Dutch-Flemish, using a standardized methodology of forward and backward translations. According to the PROMIS standards short forms can be released for use in new languages after a high quality translation.
It is expected that PROMIS instruments (especially PROMIS CATs) will soon become the gold standard world-wide for the measurement of patient-reported health. PROMIS CATs are, however, not yet available for use in the Netherlands because of two reasons:

1. CAT software and a web-based infrastructure to set up studies using CATs are not yet available. The Dutch-Flemish PROMIS group is working on this.
2. Reference scores from the Dutch general population are not yet available. These are needed because all PROMIS scores are expressed as T-scores, in which scores have a mean of 50 and a standard deviation of 10 compared with the general population.

Dutch reference scores are to be preferred over US reference scores. The implementation of the PROMIS-29 Profile in Lifelines would help to estimate Dutch reference scores for seven PROMIS item banks (if necessary, using CBS census data for weighting), which will facilitate the implementation of CATs for these banks.

Content validity of the PROMIS instruments can be considered good because they are based on a clear consensus-based conceptual model and conceptual definitions of all domains, extensive literature search, including consideration of all items from all existing Patient Reported Outcomes, qualitative research, including 79 focus groups with patients from different disease populations, and sophisticated psychometric analyses (Item Response Theory).
Construct validity was supported by moderate to strong correlations with legacy measures (i.e. traditional Patient Reported Outcomes that measure similar constructs) in the general US population. For example, moderate to strong correlations with legacy measures were found for Physical Function (0.88 with HAQ, 0.90 with SF36-PF), Fatigue (0.91 with FACIT, 0.85 with SF-36-VT), Pain Interference (0.84 with BDI Interference), Depression (0.83 with CES-D), and Anxiety (0.80 with MASQ).
Reliability of PROMIS instruments was assessed in the general US population by calculating each item bank’s standard error and reliability coefficients, by T-scores. Reliability (defined here as measurement precision along the continuum) was high for all PROMIS instruments from scores at the mean to 2 or more SD units worse than the mean.
Several studies have shown that the measurement error of PROMIS instruments is smaller than the measurement error of questionnaires measuring comparable constructs (for example PROMIS versus SF-36). These studies were performed in patients with arthritis, depression and patients in a rehabilitation setting.

In 2022, the measurement properties of the Dutch PROMIS-29 v2.1 profile was assessed in Lifelines participants with and without chronic conditions¹⁾