Emoqol-100: Development and validation of a single question for low mood in primary care. A retrospective audit.

Background Patients with depression need to be diagnosed and managed effectively in primary care. However, current inventories for case-finding low mood are time-consuming when considering the limited time available during appointments. Aim To validate the diagnostic accuracy of a single question on the emotional quality of life (Emoqol-100) as a measure of depression in symptomatic patients. Design & setting A retrospective clinical audit, validating the Emoqol-100 compared with the 9-item Patient Health Questionnaire (PHQ-9) and Burns Depression Scale Today (BDST) in South Auckland, New Zealand. Method Consecutive patients with suspected low mood, seen over 22 months in a single primary care clinic by one of the authors, were eligible for this retrospective audit (n = 160). The index test was the verbally asked Emoqol-100: 'How is your emotional quality of life now, with 100 being perfect and 0 being the worst imaginable?' The reference standard was the PHQ-9 (n = 426 visits) with a cut-off point of ≥10 or BDST (n = 513 visits) with a cut-off point of ≥6. Results The Emoqol-100 range 0–20 had a likelihood ratio (LR) of 25.2 for low mood compared with the BDST as the reference standard; and for Emoqol-100 scores of 21–40, 41–60, 61–80, and 81–100 the LRs were 3.6, 1.7, 0.35, and 0.09, respectively. For the PHQ-9, these were 10.1, 2.9, 1.3, 0.40, and 0.2, respectively. Any score ≤60 was associated with a low mood. Conclusion The Emoqol-100 appears to have high validity, so when it is low (≤60), it is suggestive of a high PHQ-9 or BDST score, and a mood issue probably exists. Emoqol-100 could be helpful for busy primary care professionals and other clinicians.


Introduction
Depression is usually managed in the primary care setting, 1,2 and the ability of primary care clinicians to effectively diagnose and manage depression is critically important. 3A meta-analysis indicated that GPs fail to diagnose more than 50% of patients with depression in their clinic, even if the diagnostic accuracy improves when the GPs meet their patients over an extended period. 4Those presenting with somatic symptoms, for which no apparent cause can be found, are less likely to be recognised than a similar group who present with depressive symptoms. 4A broad case-finding approach using a short mood inventory test can help GPs correctly identify patients with depression promptly in primary care. 5epression inventories are useful in primary care to help clinicians to determine the likelihood and degree of depressive symptoms. 6,7The PHQ-9 has been identified as one of the most valid. 8,9It relates to the patient's symptoms over the previous 2 weeks.
The Brief Mood Survey, developed by Burns, is another tool that includes three 5-item subscales for assessment of depression, anxiety, and anger during the previous 1-week period.The Brief Mood Survey has been shown to be reliable, with excellent internal consistency. 10There is also another shorter Burns questionnaire with five questions about the mood 'today' (on the day of administering the questionnaire), which the authors have labelled Burns Depression Scale Today (BDST) to make clear the distinction from the 1-week Brief Mood Survey.The BDST (Supplementary Table S1) can be used to assess low mood today and, to the authors' knowledge, it has not been validated apart from one conference abstract. 11he need for a quicker and reliable test to evaluate the severity of symptoms of low mood in primary care was identified by one of the investigators.The Emoqol-100 is a single question with a derivation validation conducted against the PHQ-9; 12 the Emoqol-100 question is as follows: 'How is your emotional quality of life now, with 100 being perfect and 0 being the worst imaginable?' The answer is scored verbally from 0 to 100.The Emoqol-100 question is verbally administered, takes <15 seconds to apply and interpret, and appears to be well understood by most patients. 12he Emoqol-100 was first used in the clinic alongside the PHQ-9.During the COVID-19 pandemic, the BDST was initiated for phone consultations, in addition to the Emoqol-100, as it was easier to administer than the PHQ-9.When time was short, only the BDST was used owing to the time needed for the PHQ-9.The PHQ-9 was also administered where possible as it was required for payment purposes, allowing comparison of the three scores.
The first study to investigate the value of a test is called the derivation study, while the second validation study should validate the derivation findings in a different population.The studies should be performed according to the STARD (Standards for Reporting of Diagnostic Accuracy Studies) statement, 13 and the final test of a diagnostic test or model should investigate whether it is accurate and generalisable enough for the purpose for which it was derived. 14Criterion validity involves comparison with a gold standard, which is called concurrent validity when the comparison is made simultaneously. 15These validation and/or derivation studies are ideally done in the settings where the diagnostic test will be used. 16his article aimed to validate the findings in the derivation study of Emoqol-100 and PHQ-9 by validating the Emoqol-100 against PHQ-9 as a reference standard, and to derivate the Emoqol-100 against the BDST questionnaire.

Method
A retrospective audit was conducted over 22 months, from 25 November 2002-28 September 2022, at a general practice clinic in South Auckland, New Zealand.Participants were consecutive patients (n = 160) seen by one of the authors, in whom low mood was a key issue and who were coming for fully funded wellness visits.A patient requests a wellness visit for a 30-minute consultation, usually for emotional distress.The clinic gets paid an increased fee for the service by the health system.Some of these patients were regular, and clinic colleagues referred others for a Focused Acceptance and Commitment Therapy (FACT) consultation. 17Patients were eligible for the audit if they had a recorded Emoqol-100 score and a PHQ-9 score or BDST questionnaire administered at the same visit.These were assessed during the visit by the GP.The Emoqol-100 was the index test, and the BDST questionnaire or PHQ-9 was the reference standard.The order of doing the Emoqol-100, the BDST questionnaire, and the PHQ-9 were variable, but the Emoqol-100 was generally done first as it was the quickest to complete and on some occasions, the only one done.The reference tests (PHQ-9 and BDST) were not given blindly to the patients.While the clinician had other information, such as medication and medical history, this did not alter the administration of the Emoqol-100 test, BDST, or the PHQ-9.Only patients with reasonably good English language abilities were asked the Emoqol-100.
The analysis was done according to the method of Guyatt and Rennie for calculating LRs.For each level of the Emoqol-100, the true positive number is divided by the total depressed and the false positive divided by the total not depressed (TP/all depressed) / (FP/those not depressed). 16A likelihood ratio >1 increases the post-test probability of the condition, while a likelihood ratio <1 decreases the post-test probability of the condition.
A recent meta-analysis reported that a PHQ-9 score of ≥10 is the level where the combination of sensitivity and specificity is maximised overall, and this was the cut-off used. 9For BDST, a score ≥6 is classified as depression and was used as a cut-off. 18The number of patients available determined the sample size during the study period.There was no public or patient involvement in this work.

Baseline characteristics
There were 160 patients and n = 426 visit records of PHQ-9 and n = 513 visit records of BDST.The findings are shown in Figures 1 and 2. The majority of the patients (62%) were women, and the median age was 35 years (14-78), as shown in Table 1.The distribution of ethnic groups reflected the general population of the clinical study site reasonably well.The median score of Emoqol-100 was 55.The median for PHQ-9 was 13, which is equal to mild-to-moderate depression, and the median for BDST was 8, equivalent to moderate-to-low mood.
Low mood was present in 69% of the sample according to PHQ-9, and in 67% according to BDST.The practice had 5000 registered patients.There were seven GPs, one GP trainee, two nurse practitioners who work as GPs four nurses, and two healthcare assistants.The practice is called a very low-cost access (VLCA) practice, meaning that more than half the patients are either from Maori or Pacific ethnic groups, or live in the most socioeconomically deprived quintile.Because it is a VLCA practice, the clinic gets more funding from the health system.

Emoqol-100 validation against PHQ-9
For patients with an Emoqol-100 score of ≤20, the LR was 10.1, with a positive PHQ-9 (≥10) as the reference standard (Table 2).The Emoqol-100 score 21-40 had an LR of 2.9, score 41-60 had an LR of 1.3, score 61-80 had an LR of 0.40, and a score of 81-100 had an LR of 0.2 (Table 2).Based  on the PHQ-9 ≥10, the highest positive predictive value was 96% for a cut-off point of ≤20 (Table 2).This means that a patient who scores ≤20 is 96% likely to have a PHQ-9 score of ≥10, indicating a high probability of a low mood at that visit and a clinically significant increase from the average low mood of 69%.

Emoqol-100 derivation against Burns Depression Scale Today
For patients with an Emoqol-100 score of ≤20, the LR of low mood is 25.2, with a positive BDST (≥6) as the reference standard (Table 3).The Emoqol-100 score 21-40 had an LR of 3.6, score 41-60 had an LR of 1.7, score 61-80 had an LR of 0.35, and score 81-100 had an LR of 0.09 (Table 3).As the Emoqol-100 score gets higher, the LR drops, and in the higher Emoqol-100 range, the patient is much less likely to suffer from a mood disorder.Based on the BDST ≥6, the highest positive predictive value was 98% for a cut-off point of ≤20 (Table 3).This means that a patient who scores ≤20 is 98% likely to have a BDST score of ≥6, which is a clinically significant increase from the average low mood on the BDST of 67%.The same applies for a Emoqol-100 score of >80; the posttest likelihood of this is a positive predictive values of 15%, which is a clinically significant decrease from 67%.

Discussion Summary
The Emoqol-100 was validated against the PHQ-9 and the results were consistent with the previous derivation study. 12The Emoqol-100 score in the low range is associated with a high PHQ-9 and a high BDST (high scores indicate low mood).The higher the Emoqol-100 score, the lower the scores for PHQ-9 and BDST.The Emoqol-100 score in the low range is associated with an increased risk of low mood, according to both the PHQ-9 and the BDST.For an Emoqol-100 score of 0-20, the LR for PHQ-9 is 10.1 and for BDST is 25.2, which is very high and suggests that the test will significantly change post-test probabilities from pre-tests.It is unusual to have such high LRs in clinical medicine. 19In addition, the Emoqol-100 with scores 81-100 have very low LRs, and a high score (a negative result) could be used to rule out low mood.

Strengths and limitations
This is a clinical audit and, as such, is a pragmatic study applied to consecutive patients with known  or suspected depression.The advantage of an audit is that it is possible to measure inventories while in use in regular clinical practice, without the confounding issues of consent and information sheets, which can cause a selection bias in planned research.The Emoqol-100 has been validated and derivated in the clinical setting where it is intended to be used.The prevalence of low mood was significantly higher than in a consecutive series of patients seen in a usual general practice setting.This sample of included patients was seen by one of the authors for extra mental health care.This should not affect the properties of the test, but rather the interpretation of a negative test since a diagnosis is hard to rule out in a highly prevalent condition.This study has several limitations.The clinical audit only applied to one practitioner, and it was impossible to blind the measurement of the reference standard PHQ-9 or BDST.The PHQ-9 is becoming the common tool for primary care assessments of depression. 20The Emoqol-100 measures the mood right now, and that can be considered both a weakness and a strength.It may be less valid than tools measuring more extended periods of low mood, but may potentially be more precise and faster to complete, as well as not being subject to recall bias.

Comparison with existing literature
Low mood was present in the cohort for 69% for PHQ-9 and 67% for BDST.This is a very high prevalence, and thereby pre-test probability is high compared with the average primary care prevalence of mood disorders of approximately 12.9%. 21The prevalence was expected to be high as the patients were selected for extra attention for their mental health.The prevalence should not affect the LR results since it is a function of the sensitivity and specificity. 22Diagnostic properties like sensitivity and specificity are not

Figure 2
Figure 2 Flow diagram for records of reference standard Burns Depression Scale Today (BDST) and Emoqol-100 score (n = 513 visits)

Table 3
Derivation assessment of Emoqol-100 with reference standard Burns Depression Scale Today (BDST)