Check each claim against its experiments, reported results, scope, and stated limitations.Know what the science actually supports before you trust the answer.
Press Enter β΅ to start review
Explore by Goal
"The first principle is that you must not fool yourself β and you are the easiest person to fool."
- Richard Feynman
Quick Answer
Copied
What this paper finds (children, German statutory insurance, 2009)
Prevalence: psoriasis 0.45% (0.42β0.47) vs atopic eczema 10.35% (10.24β10.47) among insured youth .
Comorbidity patterns differ: psoriasis shows higher metabolic diagnoses (e.g., obesity, hyperlipidaemia, hypertension, diabetes) and markedly elevated inflammatory/ocular diagnoses (e.g., arthritis and iridocyclitis), while atopic eczema shows stronger atopic-linked comorbidities (allergic rhinitis, asthma) and elevated impetigo/vitiligo .
Long Answer
Paper Review: Epidemiology and Comorbidity in Children with Psoriasis and Atopic Eczema
Journal/Year: Dermatology, 2015 .
Core evidence type: Population-based observational analysis of administrative claims using ICD-10 codes (Germany) .
Known vs inferred vs uncertain (epistemic hygiene)
Known from this paper: reported prevalences and prevalence ratios based on ICD-10-coded diagnoses in a defined insured population in 2009 .
Inferred (directional only): the authors interpret patterns as clinically meaningful comorbidity clustering that may warrant early detection and interdisciplinary management .
Uncertain: whether coded comorbidities represent true clinical diagnoses and whether healthcare-seeking differences fully explain associations; the paper explicitly flags ICD-10 coding reliability and the fact that prevalence reflects those who seek/receive care .
Figure 1 β Prevalence by age (two points per condition, paper-reported)
Values are taken directly from the Results narrative (0β2 and 14β18 age-group prevalence estimates) .
Figure 2 β Most frequent comorbidities: prevalence (%) (paper-reported)
The psoriasis-group prevalence values for allergic rhinitis, bronchial asthma, and obesity (and metabolic diagnoses) are reported in the Results; atopic-eczema prevalence values for allergic rhinitis, bronchial asthma, impetigo, and vitiligo are also reported .
Figure 3 β Diagnostic strength signals: prevalence ratios (PR) where paper reports large elevations
PRs shown are taken from the paperβs Table 1 and/or Results text: arthritis (PRβ15.25), iridocyclitis (PRβ9.55) for psoriasis; and for atopic eczema, impetigo (PRβ3.42), vitiligo (PRβ3.26), bronchial asthma (PRβ3.05), allergic rhinitis (PRβ3.04) .
Table β Paperβs PR table subset (psoriasis vs no psoriasis; AE vs no AE)
Diagnosis
Psoriasis: prevalence (%)
Psoriasis: PR (95% CI)
Atopic eczema: prevalence (%)
AE: PR (95% CI)
Arthritis
0.15
15.25 (3.64β63.98)
0.01
2.00 (0.57β6.97)
Iridocyclitis
0.38
9.55 (3.90β23.42)
0.06
2.00 (1.20β3.33)
Obesity
7.08
1.89 (1.53β2.34)
4.11
1.34 (1.26β1.42)
Hyperlipidaemia
1.14
1.79 (1.08β2.99)
0.71
1.29 (1.12β1.49)
Arterial hypertension
0.91
2.09 (1.18β3.69)
0.40
1.03 (0.85β1.24)
Diabetes mellitus
0.61
1.97 (0.98β3.96)
0.34
1.31 (1.06β1.61)
Allergic rhinitis
15.16
1.66 (1.43β1.93)
19.64
3.04 (2.95β3.14)
Bronchial asthma
12.19
1.34 (1.14β1.59)
19.04
3.05 (2.95β3.15)
Impetigo
3.35
2.32 (1.72β3.14)
3.57
3.42 (3.19β3.68)
Vitiligo
0.23
2.08 (0.67β6.49)
0.26
3.26 (2.51β4.22)
Table entries correspond to the paperβs Table 1 for selected diagnoses and are limited to rows provided in the extracted full text .
Methods critique (what could distort associations)
Case definition via ICD-10 codes: Exposure (psoriasis L40; atopic eczema L20) and comorbidities are identified by ICD-10 diagnoses; the Discussion explicitly states reliability and severity of diagnosis coding are still not clear and that diagnoses were not verified .
Surveillance/healthcare utilization bias: The authors argue patterns differ enough to suggest overreporting due solely to more physician visits is unlikely; however, residual confounding by healthcare access/frequency remains possible .
Prevalence, not incidence: The study computes prevalence in 2009; without timing of comorbidity relative to skin-disease onset, causality cannot be established (and the paper does not claim causality) .
Results critique (magnitude vs uncertainty)
Sample size is large (1.64 million total; 293,181 children/young adults; psoriasis n=1,313; AE n=30,354), which supports precision of prevalence estimates .
PR confidence intervals are wide for rare comorbidities (e.g., arthritis in psoriasis has a broad CI), indicating instability likely driven by small baseline prevalence and/or coding variability .
Comorbidity lists are ICD-10 constrained: The set of included comorbidity categories may omit relevant phenotypes or downstream clinical manifestations, so βpattern differencesβ depend on the chosen code list .
Conflict of interest & sponsorship skepticism
This paper reports research funding and speaker/consultant relationships for some authors with Biogen Idec, and one author disclosed consulting/presentations .
Because this is administrative-data epidemiology, sponsorship could plausibly influence interpretation framing (e.g., emphasis on clinical actionability) even if it is less likely to directly alter raw coding-derived prevalence ratios; the association estimates themselves remain what the data encode .
What information would most likely disprove/change the paperβs conclusions?
Validation failures for ICD-10 coding: if independent chart-level validation shows systematic misclassification differs by disease group (psoriasis vs eczema) and specifically for the high-PR comorbidities .
Healthcare utilization fully explains associations: if after stronger adjustment for contact frequency and other utilization metrics the PR pattern collapses (especially for metabolic and ocular/inflammatory diagnoses) .
The study provides an epidemiologic βfingerprintβ that psoriasis (in children) is accompanied by higher prevalence of metabolic and select inflammatory/ocular diagnoses, while atopic eczema is accompanied by higher prevalence of atopic comorbidities (rhinitis/asthma) and also impetigo/vitiligo .
The evidence is association-only and code-based; therefore, it is best treated as a hypothesis-generating comorbidity map rather than causal proof .
Moderately novel for its time: it directly contrasts pediatric psoriasis vs atopic eczema comorbidity profiles using a large, single-country administrative dataset, but the overall comorbidity concept and adult parallels were already established.
Scientific Quality
70%
Strengths include very large sample size and explicit PR/CIs. Key limitations are acknowledged: reliance on ICD-10 coding without clinical verification, possible misclassification and healthcare-utilization bias, prevalence (not incidence) design limiting temporality, and constrained comorbidity code lists. Reported sponsorship/conflicts also require skeptical interpretation of framing, though core quantitative outputs are code-derived.
Study Generality
60%
Findings are plausibly informative for pediatric comorbidity mapping in settings with similar coding/insurance structures, but generalizability beyond insured German children is uncertain because results depend on ICD-10 coding practices and healthcare-seeking patterns in that system.
Study Usefulness
70%
Useful as a hypothesis-generating epidemiologic map: it highlights differential comorbidity βsignaturesβ that can guide future validated studies. It is less useful for mechanistic causal inference or for clinical decision-making without validation studies.
Study Reproducibility
60%
Reproducibility is moderate: the paper describes the dataset source, ICD-10 code strategy, and statistical approach. However, no public accession of the dataset is provided, and diagnosis-code validity in this specific dataset is not fully quantified, limiting exact replication.
Explanatory Depth
40%
The paper is primarily descriptive/epidemiologic; mechanistic explanations are largely inferential and not tested with mechanistic data in the study design.
We'll email you the results when your analysis is finished.
Hypothesis Graveyard
A single generic explanationββmore doctor visits causes more coded diagnosesββis unlikely to fully explain the differing comorbidity spectra, but it could still reduce effect sizes substantially after stronger utilization adjustment (paper acknowledges this uncertainty).
Comorbidity differences driven only by symptom visibility/recording frequency (without disease-specific biology) is a plausible partial contributor; however, the disease-specific PR pattern and different age trajectories argue against this being the entire story (still unproven).
Quick Answer
Long Answer
Top Data Sources
ExportMCP
Hypothesis Graveyard
Potential Experiments
Discussion
Explore by Goal
Ask a Follow-Up
Key Insight
Keep Exploring