RESEARCH DISCIPLINE
Depression and Anxiety Outcome Analysis
Cutoffs, Change and Repeated Scores
Scores from the PHQ-9, the GAD-7 and similar scales, analyzed as points and as categories: severity bands applied by the published rule, response defined before outcomes are seen, and change across visits estimated with its interval.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
What Does Symptom Score Analysis Involve?
Depression and anxiety outcome analysis estimates how symptom scores differ between groups, how they change across assessments and how many people cross a defined threshold. Analysis of these scores rests on the published scoring rule for the PHQ-9 and GAD-7, since both are screening instruments with totals of 0 to 27 and 0 to 21, and on two early choices: a score or a category, and a model for repeated visits. Symptom score change analysis is reported in points with a confidence interval, set against a criterion for change stated in advance. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, assigns this work to in-house PhD-level statisticians and records every scoring and response definition beside the result it produced. Clinical interpretation is yours.
WHEN TO ENGAGE
Symptom Data Waiting for a Plan
Each situation below changes which symptom outcome leads the report and how missed visits are treated.
Score or Category to Choose The protocol asks for the mean score and for the proportion above a threshold, and the plan must name the primary outcome and state what dichotomizing a score gives up.
Response and Remission to Define Success is to be reported as response or remission, and the definitions, the time point and the treatment of missing assessments have to be fixed before outcomes are examined.
Assessments Across Weeks The same participants completed the scales at baseline and at later visits, and the questions are at what rate scores fell, for whom, and what to do about visits that were missed.
A Change Said to Matter Scores fell after an intervention, and the report must show whether the fall is large enough to count for patients, using a criterion set apart from the p-value.
SCORES AND CHANGE
From Item Responses to Reported Change
A total, a band and a change from baseline answer different questions, so each is defined separately and the primary one is fixed in advance.
Versions and Short Forms
The PHQ-8 omits the self-harm item and totals 0 to 24, and the PHQ-2 and GAD-2 total 0 to 6, so the version is confirmed before a published band or cutoff is applied.
Thresholds and Bands
Scores above a cutoff are reported as screen positive, not as diagnosed, and the proportion with its interval is accompanied by the analysis of the score the cutoff compresses.
Response and Remission
Definitions such as a halving of the baseline score are fixed in advance, and groups are compared as a risk difference or odds ratio with a 95% confidence interval.
Change Across Visits
Scores at several visits are modeled with Linear Mixed Models (Multilevel Modeling), so a participant who missed a visit still contributes, and the visit-by-group pattern is estimated as planned.
Criterion for Change
Change is judged against a criterion fixed beforehand: a reliable change index for whether one person's change exceeds measurement error, or an anchor-based threshold for whether the change is one patients notice.
Reporting by Visit
Reporting follows Counseling & Mental Health Psychology and Clinical Psychology, with the number assessed at each visit.
COMMON ISSUES
Screening Totals Read Too Literally
The errors that matter most come from treating a screening total as a diagnosis or a cutoff as a measurement, and each can be avoided in the plan.
Screen Positive Read as Prevalence The share scoring 10 or more is reported as the share with major depression, though screen positives include false alarms and omit missed cases in numbers that depend on the setting.
Cutoff Used as the Outcome Splitting a 0 to 27 score at one cutoff discards information, and a small shift in scores near the line can look like a large change in the proportion above it.
Zero-Heavy Distributions In community samples many people score near zero and the rest trail to the right, so a normal-theory model may fit badly, and the shape is plotted before a model is chosen.
Response Defined After the Fact A response threshold chosen after seeing which one favors the treatment inflates the result, so the definition, the time point and the dropouts are fixed in the plan.
Last Score Carried Forward Carrying a last observed score forward assumes symptoms stopped changing when a person left, which is often untenable, so every observed visit is modeled and the assumption is probed.
Two Scales, One Condition Depression and anxiety scores correlate substantially, so analyzing them as unrelated outcomes overstates how much independent evidence two results provide.
DELIVERABLES
Scored Data, Models and Definitions
Each document ties a number to its source: which cutoff, which response rule and which assumption about missed visits lies behind every reported estimate.
Scored Symptom File A dataset with each scale total, band and item-missing flag derived from the raw responses by the published rule, with a codebook of every derived variable.
Floor and Band Profile The shape of each total at every visit, with the pile-up near zero and the count in each severity band, which decides whether a model for skewed scores is needed.
Response and Remission Results The agreed definitions with the number meeting each at every time point, the group difference with a 95% confidence interval and the count with missing assessments.
Visit-by-Visit Estimates Mean scores and group differences at each assessment from the model, with 95% confidence intervals, the number assessed at each visit and a course figure.
Criterion Statement A note naming the criterion for meaningful change, where it comes from, and the share of participants who met it, for you to check and incorporate in your methods section.
Scoring Script, Item to Estimate One documented script runs from the raw item columns through scoring, the response rule and the visit model, so a corrected item value updates every reported estimate.
ILLUSTRATIVE EXAMPLE
Screening Scores From Pregnancy to Postpartum
An illustration of how two versions of a depression screener are combined and compared within the same women.
ILLUSTRATIVE EXAMPLE · PERINATAL DEPRESSION SCREENING RESEARCH · UNITED STATES
A nurse-midwifery researcher used depression and anxiety screening scores recorded at three obstetric clinics and asked whether symptoms and the share screening positive differed between the third trimester and the postpartum visit.
Two clinics recorded the PHQ-9 and one the PHQ-8, which omits the self-harm item, so every woman was scored on the eight shared items and the published PHQ-8 cutoff was applied to that total. Because the same women were scored twice, the change in mean score and the change in the share screening positive were estimated as paired comparisons, with the GAD-7 reported beside them rather than as separate evidence. The complication was that a positive screen in the third trimester often led to referral, so postpartum results were reported separately for women who had and had not screened positive. Screen positives were described as such, not as cases of depression, and the change was worded as a difference between two points in the perinatal period, not as an effect of childbirth or of care.
Outcome: The researcher received the paired changes in score and in the share screening positive with 95% confidence intervals, the version note, the scored dataset, the tables and the code.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
SOFTWARE
Scoring Scales and Fitting Visit Models
The choice depends on how the scales were collected and what the visit data need, and scoring is scripted so that it is applied identically at every assessment.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCH RESPONSIBILITY
Clinical Meaning Belongs to the Research Team
A PHQ-9 or GAD-7 result gains its clinical meaning from the researcher, who also chooses the instruments and draws the conclusions.
Agreed Statistical Support OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, defined in a scope that is confirmed before any work begins.
No Intellectual Contribution OnlineSPSS does not claim any intellectual contribution to your research. Authorship of the article, thesis or clinical report remains yours.
Your Research Decisions The choice of instruments, the clinical interpretation of the findings, the discussion and the conclusions remain yours, with overall responsibility.
Confidential Handling Research data and project details are handled confidentially, project information is not published, and an NDA is available where required.
FREQUENTLY ASKED QUESTIONS
PHQ-9 and GAD-7 Results Under Review
Screening thresholds, dichotomized scores and missed visits often draw objections when symptom score analyses are reviewed.
What does depression and anxiety outcome analysis cover?
Depression and anxiety outcome analysis covers the scoring of the PHQ-9, the GAD-7 and similar scales, the analysis of scores and of threshold-based outcomes, and the estimation of change across assessments with confidence intervals. Trial, clinic and community survey data all qualify, whether each person was assessed once or at many visits.
Does a PHQ-9 score of 10 or more mean a person has depression?
No. A total at or above the commonly used threshold of 10 marks a person for further assessment, and it is not a diagnosis. The share above the threshold is therefore reported as screen positive, and how closely it matches clinically assessed depression depends on the sample and the setting in which the questionnaire was used.
Should depression scores be analyzed as a score or as a category?
The score is usually the stronger primary outcome, since cutting it into categories discards information and typically lowers power. A category such as response or remission suits the question when the clinical decision is itself a threshold. Either can be primary, and the plan names which one before outcomes are seen.
Is carrying the last score forward acceptable for missed assessments?
It is rarely the best choice. It assumes that symptoms stopped changing when the person left, which is often untenable and can bias the estimate in either direction. A model that uses every observed visit makes a weaker assumption about the missing scores, and a sensitivity analysis shows how far the conclusion depends on it.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the scale versions, the visit schedule and the dataset, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Choose the Primary Symptom
Outcome With a Statistician
Whether a score, a band or a response rate should lead your report is a decision best made before the data are analyzed, and it shapes every table that follows.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014