🛠️ These articles include high-priority cards and their logic. If you notice a card you use is missing, please be assured we are working our way through them.

📖 Learning: This card finds all the wellbeing tracker records relevant to the MYCaW framework, filtering by trackers tagged to cases that are attached to clients within the viewing organisation. It filters for well-being trackers created within the date range selected. There is no filter applied for only clients created within the same date range.

We then separate the trackers by baseline and follow-up records. For each concern, we find the average score across all of the baseline records for all clients. This is the average baseline. We then do the same for each concern for the follow up records.

The average change is simply calculated by subtracting the average follow-up value from the average baseline value.

The number of clients is calculated by finding the number of distinct client IDs across all of the relevant records.

📖 Learning: This card finds all the wellbeing tracker records relevant to the MYCaW framework, filtering by trackers tagged to cases that are attached to clients within the viewing organisation. It filters for wellbeing trackers created within the date range selected. There is no filter applied for only clients created within the same date range.

We then separate the trackers by baseline and follow-up records. We find the average wellbeing score across all of the baseline records for all clients. This is the average baseline. We then do the same for the wellbeing score for the follow-up records.

The average change is simply calculated by subtracting the average follow up value from the average baseline value.

The number of clients is calculated by finding the number of distinct client IDs across all of the relevant records.

📖 Learning: This card finds all the wellbeing tracker records relevant to the ONS4 framework, filtering by trackers tagged to cases that are attached to clients within the viewing organisation. It filters for wellbeing trackers created within the date range selected.

We then separate the trackers by baseline and follow up records. We find the average wellbeing score across all of the baseline records for all clients. This is the average baseline. We then do the same for the wellbeing score for the follow up records.

The average change is simply calculated by subtracting the average follow up value from the average baseline value.

The number of clients is calculated by finding the number of distinct client IDs across all of the relevant records

📌 Note: for this card, there IS a filter applied for only clients created within the same date range.

📖 Learning: This card finds all the wellbeing tracker records relevant to the ONS4 framework, filtering by trackers tagged to cases that are attached to clients within the viewing organisation. It filters for wellbeing trackers created within the date range selected.

We then separate the trackers by baseline and follow up records. We find the average wellbeing score across all of the baseline records for all clients. This is the average baseline. We then do the same for the wellbeing score for the follow up records. For the anxiety metric in this wellbeing framework, the scale here is reversed, so a higher score means less anxiety.

The average change is simply calculated by subtracting the average follow up value from the average baseline value.

The number of clients is calculated by finding the number of distinct client IDs across all of the relevant records.

📌 Note: for this card there IS a filter applied for only clients created within the same date range.

📖 Learning

How It’s Calculated
• We measure GP appointments three months before a patient is referred to a social prescriber and three months after they are successfully discharged.
​• Example:
• A patient is referred on 1st January.
• The three-month pre-referral period is October–December.
• The patient is with the social prescriber January–February.
• They are discharged at the end of February.
• The three-month post-discharge period is March–May.
​
​Counting GP Appointments
• We track the number of times a medical record is updated by a GP or similar job titles (e.g., doctor, general practitioner).
• We prevent double-counting by ensuring that multiple updates by the same person on the same day count as one appointment.
​
​Reliability & Limitations
• This is a strong proxy measure, but not a forensic audit.
• If you compared it to an actual appointment book in EMIS/SystmOne or Joy, there might be minor discrepancies.
• However, the order of magnitude is reliable since the same method is used before and after referral.
• We use a three-month window instead of six months or a year because:
• Longer waiting lists and consultation durations mean data takes time to accumulate.
• A six-month window led to insufficient GP data for early business case evaluations.
​
​Sampling Approach
• Not all patients are included due to technical constraints.
• Pulling full medical records can strain EMIS/SystemOne’s backend servers.
• We have agreements with EMIS and SystemOne that throttle the data collection process.
• Instead, we use a statistically valid sample size:
• 30+ patients is the minimum required to draw meaningful insights.
• Beyond 30, additional data points have diminishing impact on results.
• Most of our customers have hundreds or thousands of data points, providing robust analysis.
​
​Why This Matters
• This method gives reliable data for business cases and ROI calculations.
• While not an exact appointment record, it offers a consistent, scalable way to quantify the impact of social prescribing on GP workload.

📖 Learning

How It’s Calculated
• We measure A&E attendances three months before a patient is referred to a social prescriber and three months after they are successfully discharged.
• Example:
 • A patient is referred on 1st January.
 • The three-month pre-referral period is October–December.
 • The patient is with the social prescriber January–February.
 • They are discharged at the end of February.
 • The three-month post-discharge period is March–May.

Counting A&E Attendances
• We analyse medical records for mentions of A&E attendance or related discharge documentation, including discharge summaries and letters.
• These are reliable indicators of an A&E visit, as they are generated only when a patient has attended or been discharged from an emergency department.
• To prevent inflated counts caused by duplicate or erroneous coding, we apply a three-day smoothing rule. Multiple A&E-related entries for the same patient within a three-day period count as one attendance.
• This helps remove data artefacts from partial or repeated record updates while maintaining accurate representation of real attendances.

Reliability & Limitations
• The measure provides a strong proxy for A&E attendance but is not a forensic audit.
• Some attendances may not generate a coded summary immediately, and occasional duplicates or missing data may occur depending on the clinical system’s configuration.
• Because the same method is applied before and after referral, any small inconsistencies affect both periods equally, preserving the reliability of the change observed.
• The three-month window balances statistical robustness with timeliness. Longer windows delay feedback and are affected by data latency in clinical systems.

Sampling Approach
• As with the GP appointment methodology, not all patients are included.
• Pulling full medical records for every patient can overload EMIS or SystmOne servers, so the process is throttled in agreement with system providers.
• We therefore use a statistically valid sample.
• A minimum of 30 patients provides meaningful insight; beyond that, additional data points have diminishing impact.
• Most analyses involve hundreds or thousands of patients, ensuring robust findings.

Example
If a patient attended A&E on 4th March and again on 6th March, this is counted as one attendance within the three-day window. A later visit on 12th March would be counted separately.