It does not require insightful analysis to conclude that something is broken when it comes to reporting and addressing sexual misconduct in the workplace.

One attempt to fix part of the “brokenness” comes from the Pennsylvania legislature – a bill that would place limits on confidentiality restrictions. (New Jersey is considering something similar and other states, such as California, are assessing possible revisions to reporting and investigative processes).

The proposed law specifically permits two pieces of information to remain confidential – the name of the person making a complaint and the monetary amount of any settlement. But it would ban portions of agreements, including settlements, and prohibit attempts to enforce provisions that do any of the following:

• Prohibit disclosure of the name of someone suspected of sexual misconduct

• Tries to conceal information that pertains to an investigation of sexual misconduct

• Restricts or impairs the ability of someone to report a concern of sexual misconduct

• Waives rights or remedies of someone relating to sexual misconduct claims

• Requires information about sexual misconduct claims to be removed or expunged unless an investigation has concluded the claims are false

The proposed law is directed at a positive goal: Not allowing a pattern of misconduct to be obscured or overlooked because incidents are often individually handled with settlements that require confidentiality.

But while the reasoning behind the proposed law is clear, the language in Pennsylvania’s bill raises a number of questions. For instance, does the prohibition against waiving rights or remedies mean the company cannot get a valid release of claims?  And will the requirement of only being able to expunge information based on a conclusion that the claims are “false” impact the outcome of investigations, especially if results are inconclusive or even mixed?

So that leaves you with the question of what actions might help to avoid a circumstance where patterns of sexual misconduct go unnoticed. At a minimum, consider some of these options:

• Create opportunities to see or recognize a pattern. Don’t limit an investigation to people identified as witnesses. Instead, make a point to ask coworkers about any interactions or observations that may indicate a problem.

• Preserve institutional knowledge. Do not count on individual knowledge as a way to spot a pattern. After all, individuals change positions or even companies. There should be an internal mechanism, appropriately secured and with limited access, that allows a company to determine if there are prior, similar complaints.

• Consider confidentiality carve outs. At a minimum, have an agreement that allows for appropriate follow up, such as responding to further company inquiries or in response to a legal inquiry, whether involving litigation or a government agency investigation.

• Encourage continued reporting. Be sure to affirm that any future or similar incidents, along with anything that could be considered retaliatory, should be reported. And then check in, and give the reminder again.

• Start out by protecting all of the individuals involved. Legally, companies are not required to be the manners police. And despite all the recent public disclosures apparently revealing numerous ignored patterns of bad behavior, some things labeled as “harassment” or “sexual misconduct” represent more mundane personality conflicts. Many investigations have inconclusive results. No individual should be labeled or face consequences as a harasser solely based on an individual complaint.

• Don’t forget about protected, concerted activity. Keep in mind the National Labor Relations Board already protects the rights of workers to share information regarding the terms and conditions of employment (generally referenced as “protected, concerted activity”).

Finally, continue to be on the lookout – for new thoughts and best practice ideas about training, reporting, investigation, and related topics. Build processes that balance consistency and flexibility. Individuals are different; what’s optimal for one person may be marginal or even negative for another.